Post-traumatic stress disorder symptomatology among American Indian Vietnam veterans

Material Information

Post-traumatic stress disorder symptomatology among American Indian Vietnam veterans mediators and moderators of the stress-illness relationship
Dempsey, Catherine Lisle
Publication Date:
Physical Description:
vi, 266 leaves : illustrations ; 28 cm


Subjects / Keywords:
Post-traumatic stress disorder ( lcsh )
Indian veterans -- Mental health ( lcsh )
Vietnam War, 1961-1975 -- Psychological aspects ( lcsh )
Combat Disorders -- Meta-Analysis ( mesh )
Indians, North American -- psychology -- Meta-Analysis ( mesh )
Risk Factors -- Meta-Analysis ( mesh )
Stress Disorders, Post-Traumatic -- Meta-Analysis ( mesh )
Veterans -- psychology -- Meta-Analysis ( mesh )
Vietnam Conflict -- Meta-Analysis ( mesh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 243-266).
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Catherine Lisle Dempsey.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
62717751 ( OCLC )
RC552.P67 D46 2001a ( lcc )
2006 C-998 ( nlm )

Full Text
Catherine Lisle Dempsey
B.S., Emory University, 1984
M.P.H., Emory University School of Public Health, 1989
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences

2001 by Catherine Lisle Dempsey
All rights reserved.

This thesis for the Doctor of Philosophy
degree by
Catherine Lisle Dempsey
has been approved
Christina Mitchell
3, JZ60I

Dempsey, Catherine Lisle (Ph.D., Health and Behavioral Sciences)
Post-traumatic Stress Disorder Symptomatology among American Indian Vietnam
Veterans: Mediators and Moderators of the Stress-Illness Relationship
Thesis directed by Associate Professors Janette Beals and
Christina Mitchell.
Results from the National Vietnam Veterans Readjustment Study (NVVRS)
reported high rates of Post-traumatic Stress Disorder (PTSD) among Vietnam Theater
veterans compared to rates in the Vietnam Era and others of the veterans' generation.
Prevalence rates were even higher among minority groups, specifically Blacks and
Hispanics. Results from the American Indian Vietnam Veterans Project (ATWP)
suggested that American Indian Vietnam veterans were also at increased risk for
PTSD. However, not all American Indian veterans with high levels of trauma
exposure developed PTSD, which suggests that other contributing factors specific to
American Indian populations may also affect their vulnerability to PTSD outcomes.
The objective of this study was to identify potential predictors of PTSD
symptomatology across three military timeframes and to examine the
relationships among personal resources, trauma, and PTSD symptomatology in

American Indian Vietnam veterans. It was hypothesized that high levels of social
support and ethnic identity may enhance one's psychosocial resilience to stress,
resulting in positive health outcomes. This study was based on AIWP data
collected by the National Center for American Indian and Alaska Native Mental
Health Research (NCAIANMHR) at the University of Colorado Health Sciences
Interviews with 621 American Indian Vietnam veterans living on or near
their reservations assessed predisposing factors, characteristics of military service,
military and nonmilitaiy trauma, personal resources, and PTSD symptomatology.
The results of hierarchical linear regression analyses showed a strong relationship
between social support and PTSD symptomatology across all time frames.
Although results did not support the stress-buffering hypothesis, combat trauma
and social support during the military interacted significantly. In addition, post-
military social support appeared to mediate the relationship between trauma and
PTSD symptomatology. Identifying a relationship between social support and
PTSD has implications for the development of interventions used to treat PTSD in
ethnic minorities. The impact of personal resources on PTSD symptomatology
may be important for traumatic survivors and long-term strategies for victims of

This abstract accurately represents the content of the candidate's thesis.
We recommend its publication.
Christina Mitchell

I dedicate this dissertation with gratitude to my family who, with
encouragement and much support, played an important role in bringing this
project to fruition. I would also like to dedicate this dissertation to the American
Indian Vietnam veterans who served their countries and shared their stories of

This research was supported in part through grants from the National
Center for American Indian and Alaska Native Mental Health Research Center,
which is supported by the National Institute of Mental Health (NIMH: MH43175,
Spero Manson, Director) and the Veterans Administration as a supplement to
I also wish to acknowledge Drs. Janette Beals and Christina Mitchell, for
their guidance, input, and long-term commitment to this project; and Drs. Debbie
Main, Diana Gurley, and Clydette Stulp for their many contributions to this
document. I would also like to thank the American Indian communities, and
families for their extraordinary support; and the locators, the interviewers, and the
field site staff who made the AIWP data collection possible.

1. INTRODUCTION....................................................1
Specific Aims.................................................4
Description of Study Population...............................7
Overview of Research Methods..................................7
Organization of the Dissertation..............................8
2. REVIEW OF THE LITERATURE........................................9
History of the Diagnostic Criteria of PTSD....................9
Epidemiology of PTSD.........................................13
PTSD in the General Population............................14
PTSD in High-Risk Populations.............................15
PTSD and Veterans.........................................16
Stress-Vulnerability Theory..................................28
Vulnerability Factors

Mediators and Moderators of PTSD
3. METHODS.........................................................58
General Framework of the Present Study.....................58
AIVVP Research Design...................................59
AIVVP Study Population..................................60
AIVVP Sampling Strategy.................................61
AIVVP Data Collection Procedures........................65
Scale Development Procedures............................67
Outcome Variable........................................70
Predisposing Factors....................................91
Characteristics of Service during the Military..........96
Post-Military and Current Demographics..................99
Social Support Scales..................................103
Ethnic Identity Scales.................................127
Stress Construct.......................................134
4. RESULTS........................................................140
Descriptive Statistics of Variables.......................140
Missing Data

Hierarchical Multiple Regressions.............................157
Specific Aim 1.............................................157
Specific Aim 2.............................................160
Specific Aim 3.............................................163
Specific Aim 4.............................................165
Specific Aim 5.............................................167
Specific Aim 6.............................................169
Specific Aim 7.............................................178
Summary of Findings...............................................180
5. DISCUSSION..........................................................182
Review of Results.............................................182
Predisposing Factors.......................................183
Military Factors...........................................187
Post-Military Factors......................................191
Current Time Frame.........................................192
Limitations of the Study......................................197
Theoretical Limitations

Sampling Limitations...................................198
Research Design Limitations............................199
Assessment Issues......................................201
Future Directions.........................................203
Policy Implications.......................................206
Support Groups.........................................207
Informal Networks......................................208
Severity of the Stressor..................................210
Preparation for Military Stressor.........................210
Debriefing after Exposure to Military Trauma..............211
Development of Cohesive Military Units....................211
APPENDIX ............................................................214
A. SURVEY INSTRUMENT..............................................214

1.1 The Full Model Tested Across All Specific Aims...............................6
2.1 Stress-Vulnerability Theory.................................................30
4.1 Interaction between Combat Severity and
Social Support during the Military........................................171
4.2 The Final Model............................................................181

2.1 Summary of Combat-related PTSD Prevalence Estimates in Community
3.1 Northern Plains Pilot Location Results...............................63
3.2 Southwest Pilot Location Results.....................................64
3.3 AIVVP Location and Lay-Interview Results.............................65
3.4 MPTSD Scale Item Characteristics.....................................73
3.5 MPTSD Item Correlations by Tribe.....................................81
3.6 MPTSD Scale Principal Components Analyses for the Total Sample.......88
3.7 PTSD Symptomatology Second-Order Components Analyses
for the Total Sample................................................90
3.8 Summary of PTSD Symptomatology Scales................................91
3.9 Summary of Pre-military Factors......................................95
3.10 Summary of Characteristics of Military Service......................98
3.11 Summary of Post-Military and Current Demographic Factors...........102
3.12 Social Support during the Military Item Characteristics............104
3.13 Social Support during the Military Correlations by Tribe...........106
3.14 Social Support during the Military Principal Components Analyses
for the Total Sample...............................................107

3.15 Post-Military Social Support Item Characteristics...................110
3.16 Post-Military Social Support Item Correlations by Tribe.............114
3.17 Post-Military Social Support Principal Components Analyses
for the Total Sample................................................116
3.18 Post-Military Social Support Second-Order Components Analyses
for the Total Sample................................................117
3.19 Current Social Support Scale Item Characteristics...................120
3.20 Current Social Support Item Correlations by Tribe ..................122
3.21 Current Social Support Principal Components Analyses
for the Total Sample................................................124
3.22 Current Social Support Scale Items Second-Order Components Analyses for
the Total Sample....................................................125
3.23 Summary of Social Support Scales....................................126
3.24 Pre-Military Ethnic Identity Item Characteristics...................128
3.25 Current Ethnic Identity Item Characteristics........................129
3.26 Current Ethnic Identity Correlations by Tribe.......................131
3.27 Current Ethnic Identity Principal Components Analyses
for the Total Sample................................................132
3.28 Summary of Ethnic Identity Measures.................................133
3.29 Summary of Stress Construct Measures................................138
4.1 Descriptive Statistics of Variables across All Time Frames..........142
4.2 Variables with Missing Data.........................................146

4.3 Comparisons of combat severity and nonmilitary exposure for those with and
without valid data for the TEB........................................148
4.4 Correlation Analyses for the Total Sample...............................152
4.5 Independent Variable Available for Regression Analyses..................155
4.6 Specific Aim 1: Multiple Hierarchical Regression Analyses..............159
4.7 Specific Aim 2: Multiple Hierarchical Regression Analyses..............162
4.8 Specific Aim 3: Multiple Hierarchical Regression Analyses..............164
4.9 Specific Aim 4: Multiple Hierarchical Regression Analyses.............166
4.10 Specific Aim 5: Multiple Hierarchical Regression Analyses..............168
4.11 Specific Aim 6: Multiple Hierarchical Regression Analyses..............173
4.12 Specific Aim 6: Multiple Hierarchical Regression Analyses..............176
4.13 Specific Aim 8: Multiple Hierarchical Regression Analyses..............179

This chapter introduces the rationale and specific aims of this study, briefly
describes the study population and the research methods, and outlines the
organization of this dissertation.
The stress-illness relationship, as well as the individual and situational
characteristics that may make one more or less susceptible to the effects of stress,
has been studied extensively (Aldwin, 1994; Bloom, 1984; Dohrenwend, 1998;
Kessler & McLeod, 1985; Pearlin, 1989; Thoits, 1983; Turner, Wheaton, & Lloyd,
1995b; Vingerhoets & Marcelissen, 1988). The general consensus is that severe
stressors may lead to psychological distress and disorder; however, reactions to
severe stressors vary, leading to the investigation of other risk/protective factors. In
particular, the relationship between traumatic stress, risk factors, and Post-traumatic
Stress Disorder (PTSD) has been a major focus of research for the past two decades
(Foy, Sipprelle, Rueger, & Carroll, 1984; Helzer, Robins, & McEvoy, 1987;
Kessler, Sonnega, Bromet, Hughes, & Nelson, 1995; Kulka et al., 1990; Robins &
Regier, 1991; Stefansson, 1991).

Exposure to a traumatic event is a necessary requirement for a diagnosis of
PTSD (American Psychiatric Association, 1987, 1994). However, PTSD does not
occur in everyone who is exposed to a traumatic event (Kessler et al., 1994; Kessler
et al., 1995; Kulka et al., 1990; National Center for American Indian and Alaska
Native Mental Health Research, 1997). These studies have brought into question
whether or not the traumatic event itself explains PTSD outcomes and have led to
the investigation of other risk and protective factors to explain the varied reactions
to trauma.
Major known risk and protective factors for PTSD include the following:
predisposing factors (e.g., socioeconomic status, gender, ethnicity, prior traumatic
experiences, psychological characteristics) (Breslau, Chilcoat, Kessler, & Davis,
1998; Davidson, Smith, & Kudler, 1989; Egendorf, Kadushin, Laufer, Rothbart, &
Sloan, 1981; Fontana & Rosenheck, 1993; Fontana & Rosenheck, 1994; Jordan et
al., 1992; National Center for American Indian and Alaska Native Mental Health
Research, 1997); the nature and severity of the traumatic stressor (e.g., the horror
and fear it engenders, its unusualness) (Fontana & Rosenheck, 1999; Green &
Lindy, 1994; King, King, Fairbank, Keane, & Adams, 1998; Kulka et al., 1990;
Yehuda, Southwick, & Giller, 1992); and personal resources (e.g., social support,
coping, hardiness, ethnic identity) (Benotsch et al., 2000; Green & Berlin, 1987;
Keane & Fairbank, 1983; King, King, Foy, Keane, & Fairbank, 1999; Sutker,
Davis, Uddo, & Ditta, 1995b).

The protective factor most noted and described in the literature is social
support. While the results of research generally confirm that trauma severity is
related to PTSD symptoms, and that high levels of social support correspond to
lower levels of PTSD symptoms, less is known about the interrelationships between
trauma, social support, and PTSD symptoms.
This dissertation focused on American Indian Vietnam veterans, a
population with high current prevalence rates of PTSD (Beals et al., under review;
National Center for American Indian and Alaska Native Mental Health Research,
1997). Factors unique to American Indian populations such as reservation life,
discrimination, and forced acculturation may place American Indians at increased
risk for psychological distress. Therefore, this study will examine the role that
ethnic identity may have in relationship to psychological distress. Thus, the
relationships among trauma, personal resources such as social support and ethnic
identity, and PTSD symptoms among the American Indian population are still
unclear and remain to be elucidated.
Therefore, the overall goal of the present study was to examine potential
predictors of PTSD symptomatology, specifically focusing on the relationship
between severity of trauma and PTSD symptomatology among American Indian
Vietnam Theater veterans, and to determine if personal resources such as social
support and ethnic identity mediate or moderate the impact of trauma. This study
could increase our knowledge about mediators and moderators of PTSD

symptomatology and have implications for preventive/intervention programs
targeting individuals with PTSD symptomatology.
Specific Aims
Hypothesized risk and protective factors in literature focusing on veterans
and PTSD outcomes are categorized as pre-military (occurring before induction),
military (occurring during period of service), or post-military (occurring after
military service). The specific aims listed below are organized by time frame.
Figure 1 on page 6 depicts the full model to be tested.
1. In Specific Aim 1, the relationships between pre-military
predisposing factors and PTSD symptomatology will be explored.
2. In Specific Aim 2, the relationships among military characteristics,
social support during the military, severity of military trauma, and
PTSD symptomatology will be explored.
3. In Specific Aim 3, the relationships between post-military social
support and PTSD symptomatology will be examined.
4. In Specific Aim 4, the relationships among current social support,
current ethnic identity and PTSD symptomatology will be
5. In Specific Aim 5, the relationships among current demographics
and PTSD symptomatology will be explored.

6. In Specific Aim 6, the relationships between moderators and
mediators of military trauma and PTSD symptomatology will be
7. In Specific Aim 7, the full model across all specific aims will be
tested, and the existing significant relationships determined in the
previous aims will be examined.

Figure 1.1: The full model tested across all specific aims.
Control variable

Description of Study Population
This dissertation is based on the American Indian Vietnam Veterans Project
(ATVVP), for which I was the training and field coordinator for three years at the
National Center for American Indian and Alaska Native Mental Health Research
Center (NCAIANMHR), University of Colorado Health Sciences Center. The
study population consisted of male members of two American Indian tribes who
served on active duty in the U.S. Armed Forces in the Vietnam Theater1 between
August 5,1964, and May 7,1975, and who lived on or near their respective
reservations in 1993 and 1995. All participants were interviewed on their
respective reservations. The names of the reservations are not specified here to
protect the confidentiality of the communities.
Overview of the Research Methods
To accomplish the specific aims, the study employed quantitative data-analytic
methods. Hierarchical multiple linear regression methods were used to examine the
significant relationships between sets of independent variables suggested by the
Stress-Vulnerability Theory and the dependent variable (Dohrenwend &
Dohrenwend,1984; Bloom, 1984). The independent variables covered all three time

frames (pre-military, military, and post-military), encompassed areas relevant to the
literature, and meaningful to PTSD symptomatology. The dependent variable for all
analyses was PTSD symptomatology, a continuous measure of symptom severity
(Keane & Penk, 1988).
Organization of the Dissertation
The following chapter reviews the pertinent literature on Stress-
Vulnerability Theory, stress, social support, ethnic identity, and PTSD
symptomatology. Chapter 3 describes the methods and reviews the psychometric
properties of all scales created for this study. Chapter 4 explains the hierarchical
linear regression models and presents the resulting sets of factors that contributed
significantly to PTSD symptomatology across all time frames, including mediating
and moderating relationships between combat severity and PTSD symptomatology.
Chapter 5 provides the conclusions and limitations of the study, and suggests how
the research findings will contribute further to existing research. 1
1 Theater veterans refers to those veterans who served in the Vietnam Theater of operations
including those who were stationed in Vietnam, Laos, or Cambodia, stationed in waters in or
around these countries, or flew air missions over these areas.

This chapter provides a rationale for the specific aims addressed in this
dissertation. It includes an overview of the history of the diagnostic criteria of
PTSD, and a review of the epidemiology ofPTSD, which together demonstrate
gaps in the literature and needs for future research. Next, a discussion of Stress-
Vulnerability Theory provides the theoretical framework for characterizing PTSD
among American Indian Vietnam veterans. This section also introduces the
concept of stress and highlights the literature on hypothesized vulnerability factors.
Finally, this chapter concludes with a summary of mediator/moderators of the
stress-illness relationship, social support, and ethnic identity.
History of Diagnostic Criteria of PTSD
Considerable debate has focused on the diagnostic legitimacy of PTSD
(Yehuda & McFarlane, 1995). Historically, in the original Diagnostic and
Statistical Manual of Mental Disorders (DSM)(American Psychiatric Association,
1952) the first formal diagnosis of trauma-based disorders was referred to as "gross
stress reaction" in response to the compelling numbers of investigations describing
the emotional, psychological, and physical distress of World War II veterans. The

DSM explicitly stated that extraordinary extreme stressors could result in a "gross
stress reaction." This diagnosis sparked a series of studies on the etiology and
phenomenology associated with traumatic stress. In the 1970s, DSM-II broadened
the definition of post-traumatic syndromes by categorizing the effects of traumatic
stress under the diagnosis of "transient situational disturbance," if the symptoms
were short-lived, and "anxiety neurosis," if the symptoms persisted (American
Psychiatric Association, 1952).
The Vietnam War influenced the creation of a new category, Post-traumatic
Stress Disorder. In 1980, DSM-III (American Psychiatric Association, 1980)
confirmed PTSD as a major diagnostic entity within the anxiety disorders. In
addition, DSM-III outlined the disorder's primary symptoms in some detail, which
helped to expand the field of traumatic stress research. Researchers developed
reliable and valid measures of PTSD, using the DSM-III's consistent
conceptualization of PTSD diagnosis.
Identified symptom criteria included the following: Criterion A required
exposure to an extreme stressor; Criterion B required the presence of symptoms
where the traumatic event is persistently re-experienced by recurrent, intrusive
dreams, nightmares, or intense psychological distress at exposure to events that
symbolize the traumatic event; Criterion C required persistent avoidance of stimuli

associated with the trauma and general numbness (e.g., avoidance of thoughts or
feelings associated with the trauma and diminished interest or participation in
activities); and Criterion D required the presence of persistent symptoms of
increased arousal (e.g., difficulty falling asleep or irritability).
DSM-IH-Revised (III-R) further refined the existing PTSD diagnosis
(American Psychiatric Association, 1987). The major contribution of DSM-ffl-R
was clearer specification of the type of stressor that might produce PTSD symptoms
as a psychologically distressing event that was "outside the range of human usual
experience and that would be markedly distressing to almost anyone" (American
Psychiatric Association, 1987). The DSM-IH-R diagnostic criteria for PTSD are
provided below (APA, 1987, p. 250-251). 1
A. The person has experienced an event that is outside the range of usual
human experience and that would be markedly distressing to almost
B. The traumatic event is persistently re-experienced in at least one of the
following ways:
1. recurrent and intrusive, distressing recollections of the event
2. recurrent distressing dreams of the event
3. sudden acting or feeling as if the traumatic event were
recurring (including "flashback" or dissociative episodes,
whether or not intoxicated)
4. intense psychological distress at exposure to events that
symbolize or resemble an aspect of the traumatic event,
including anniversaries
C. Persistent avoidance of stimuli associated with the trauma or

numbing of general responsiveness, as indicated by at least three of
the following:
1. efforts to avoid thoughts or feeling associated with the
2. efforts to avoid activities or situations that arouse
recollections of the trauma
3. inability to recall an important aspect of the trauma
4. markedly diminished interest in significant activities
5. feeling of detachment or estrangement from others
6. restricted range of affect
7. sense of foreshortened future
D. Persistent symptoms of increased arousal (not present before the
trauma), as indicated by at least two of the following:
1. difficulty falling or staying asleep
2. irritability or outbursts of anger
3. difficulty concentrating
4. hypervigilance
5. exaggerated startle response
6. physiological activity upon exposure to events that symbolize or
resemble an aspect of the traumatic event
E. Duration of disturbance (symptoms in "C" and "D") of at least one month.
Empirical findings from the DSM-IV field trials recommended additional
revisions to the existing diagnostic criteria (Davidson & Foa, 1993; Resnick,
Kilpatrick, Dansky, Saunders, & Best, 1993). Some of the notable changes from
DSM-HI-R to DSM-IV included removing the provision that the stressor had to be
"outside the range of normal human experience." In its place, Criteria A made
modifications to show the importance of subjective perception and appraisal in
response to a traumatic event. In particular, in order to meet Criterion A, events

must engender intense fear, helplessness, or horror on the part of the victim. Other
changes included rearranging DSM-HI-R symptom clusters, adjusting diagnostic
thresholds, and addressing the longitudinal course of the disorder.
The DSM-m and DSM-DI-R criteria represent a standardized definition
reflecting current conceptualizations of PTSD that have informed two decades of
traumatic stress research; however, they also reflect the difficulties of identifying
and defining the core symptoms associated with PTSD and pose challenges for
fixture research. DSM-IV was published after data collection for this study began;
therefore, this study relies on the DSM-IH-R definition of PTSD. Specifically,
PTSD symptomatology was assessed using the Mississippi Scale for Combat-
related PTSD, which is based primarily on DSM-Hl-R criteria, with items added to
incorporate associated features of combat-related PTSD including depression,
substance abuse, and impairment in social and occupational fimctioning (Keane &
Penk, 1988).
Epidemiology of PTSD
This subsection provides an overview of the prevalence of PTSD, to highlight
gaps in the literature and to establish the need for additional research. The focus here
is on community rather than clinical studies. Clinical studies provide valuable insight
for improving our understanding of those who seek treatment for PTSD; however, the
results from clinical studies may be biased because only a small proportion of those

who meet diagnostic criteria for a disorder actually seek treatment (Shapiro et al.,
1984). Unlike clinical studies, community studies are aimed at assessing specific
exposures/disorders among a particular population, regardless of whether the
individual has sought treatment. The samples for the community studies are
representative of the population of inference (e.g., the general population, veterans,
crime victims, incest survivors). Special attention is given to the veteran population,
and more specifically the populations studied in the National Vietnam Veterans
Readjustment Study (NVVRS) and the American Indian Vietnam Veterans Project
PTSD in the General Population
Research on PTSD has consistently revealed lifetime prevalence rates of
1% to 2% in the general population using DSM-III criteria and the Diagnostic
Interview Schedule (DIS)2 (American Psychiatric Association, 1987; Helzer et al.,
1987; Robins & Regier, 1991; Stefansson, 1991). More recently, in the National
Comorbidity Survey, Kessler and his colleagues (1995) found lifetime prevalence
rates of 7.8% and 30-day prevalence rates of 2.8% among a national probability
sample using the Composite International Diagnostic Interview (CEDI) and DSM-
m-R criteria (CIDI; Kessler et al., 1995). These studies demonstrated rates of
2 The Diagnostic Interview Schedule (DIS), a structured lay-interview for diagnosing psychiatric disorders in the general
population, has demonstrated reliability (Robins & Regier, 1991).

PTSD differed by type of trauma experience. This range of reported prevalence
rates may be attributed to the instrumentation and the populations studied; however,
these rates still translate into a minimum estimate of 2 -16 million of lifetime
PTSD cases in the United States, alone.
Results of the National Comorbidity Survey also indicated that 39% of the
general population may have been exposed to at least one traumatic event in their
lifetimes, with many of these having multiple exposures (Kessler et al., 1995).
Thus, exposure to a traumatic event was not a rare occurrence, even among the
general population, and stressors of varying degrees and types may lead to PTSD.
PTSD in High-Risk Populations
Epidemiological studies have examined survivors of a range of different
types of trauma, including combat veterans (Goldberg, True, Eisen, & Henderson,
1990; Kulka et al., 1990; The Centers for Disease Control, 1988), civilian refugees
from wars (Hinton et al., 1993; Sack, McSharry, Kinney, Seely, & Lewinsohn,
1994), survivors of natural disasters (Green, Grace, Lindy, Gleser, & Leonard,
1990; McFarlane & Papay, 1992; Shore, Volmer, & Tatum, 1989), victims of crime
(Kilpatrick, Saunders, Amick-Mcmullan, & Best, 1989; Winfield, George, Swartz,
& Blazer, 1990), and survivors of child abuse (McLeer, Deblinger, Henry, &
Orvaschel, 1992; Resnick et al., 1993). As expected, the prevalence rates of PTSD
within these special populations were higher than in the general population. For

example, Shore and colleagues (1989), using the DIS and DSM-III criteria, reported
lifetime PTSD prevalence rates of PTSD of 3.6% in the population exposed to the
Mount St. Helen's volcanic eruption, compared with 2.6% in the control group.
Based on the General Health Questionnaire (Goldberg, 1972) and DSM-III criteria,
McFarlane (1988b) reported prevalence rates of 32%, 27%, and 30% obtained at 4,
11, and 29 months post-exposure to an Australian brushfire, among a sample of 315
fire fighters (Goldberg, 1972; McFarlane, 1988b). Green (1990) reported current
PTSD prevalence rates of 28% using the Psychiatric Evaluation Form (Endicott and
Spitzer, 1972) in a 12-year follow-up of 120 survivors of the 1972 flood in Buffalo
Creek, West Virginia (Endicott & Spitzer, 1972). These studies indicate that a
variety of extreme events precipitate PTSD reactions in at least a portion of exposed
PTSD and Veterans
The prevalence of PTSD has been more widely studied among veterans than
among any other at-risk group. Veteran populations that have been studied in the
last decade include veterans from World War II, Korea, the conflicts in Lebanon
and the Persian Gulf, and the Vietnam War. The studies reviewed here focus on
veterans from these wars, including only those published after PTSD diagnostic
criteria were identified. Some of the studies are problematic because of specialized
sampling, small sample sizes, low response rates, and issues surrounding

assessment of PTSD. See Table 2.1 for a listing of these prevalence studies by each
World War II. Studies of World War II veterans and prisoners of war have
provided descriptive information about the onset of symptoms and the longitudinal
course of PTSD. For example, in studies of prisoners of war from World War II,
Kluznick (1986) found current PTSD prevalence rate of 47% using a clinician-
administered PTSD symptom checklist based on DSM-in criteria. They reported
that in this sample, aside from one possible case of delayed-onset PTSD, 21% of
the sample still had moderate residual PTSD symptoms 40 years after the war
(Kluznick, Speed, & Van Valkenburg, 1986).
In another study of prisoners of war from World War II, using a self-report
symptom checklist based on DSM-HI criteria, Zeis and Dickman (1989) found
higher current PTSD prevalence rates (55%) than did the Kluznick study (Zeis &
Dickman, 1989). Recent studies have used standardized measures to assess PTSD
and reported higher current prevalence rates among prisoners of war. For example,
Sutker (1993) compared a group of World War II combat veterans to a group of
World War II prisoners of war, using the DIS administered 40 years after the war
(Robins & Regier, 1991). The author reported current prevalence rates of 18%
among the combat veterans and 70% among the prisoners of war (Sutker, Albert,
Allain, & Winstead, 1993). This study suggested that severe stressors such as
captivity can lead to high rates of chronic PTSD outcomes.

Korean War. The Korean War provided an opportunity to examine the
long-term effects of combat-related trauma and PTSD outcomes. Similar to the
World War I studies, Eberly & Engdalh (1991) found high lifetime prevalence rates
of 70 to 79 among POWS from World War II who served in Europe and who
served in the Pacific, and reported rates of 59 for POWS from the Korean war
(Eberly & Engdahl, 1991). In a study from treatment seeking veterans, Fontana and
Rosencheck (1994) compared World War II veterans, Korean veterans, and
Vietnam veterans using the Structured Clinical Interview (SCID) and DSM-III
criteria. They reported the Korean veterans were more distressed than the World
War II veterans, and the Vietnam veterans, even though the Vietnam veterans were
exposed to more combat than the Korean veterans. The authors attributed the
higher levels of symptomatology and lower level of combat exposure found in the
Korean veteran to the unpopularity of the Korean war and the lack of respect for the
Korean veteran from his home community (A. Fontana & R. A. Rosencheck, 1994).
Spiro (1994) examined the long-term effects of combat exposure on psychological
health in a large sample of community-residing World War II and Korean veterans
who participated in the Normative Aging Study (Bosse, Ekerdt, & Silbert, 1984).
Using the MPTSD Scale the authors reported higher MPTSD scale scores for the
combat-exposed Korean veteran (60.5) compared to the non-exposed veteran, 60.5,
56.2 respectively (Spiro & Schnurr, 1994). These studies demonstrated the
relationship between combat exposure and PTSD symptoms 45 years after the war.

Lebanon War. Studies of Israeli soldiers from 1982 Lebanon War shed
light on the relationship between combat stress reaction (CSR) and PTSD. CSR is
defined as a breakdown in military functioning that occurs during or immediately
after exposure to war-related trauma (Solomon, Weisenberg, & Schwartzwald,
1987a) These studies were designed to clarify if PTSD consisted of an acute
reaction component or if CSR was a predictor of PTSD. For example, Solomon
(1987) examined current prevalence rates of PTSD one year after the Lebanon
War among those veterans with and without CSR during the military. (Controls
without CSR matched demographically to those with CSR). Using the PTSD
Inventory, a self-report symptom checklist of DSM-EH criteria, Solomon found
higher prevalence rates among those veterans who had reported CSR during the
military (59%) than for veterans with no CSR (16%) (Solomon & Mikulincer,
1987b). This study demonstrated that CSR is not PTSD, although it may make
veterans more susceptible to PTSD.
Persian Gulf War. A number of studies have examined Persian Gulf War
(i.e., Operation Desert Storm) veterans. Some of these studies were problematic
because of the samples used and limits of the methods used to assess PTSD
diagnosis. For example, some of these studies used a measure of PTSD
symptomatology (e.g., the Mississippi Combat-related PTSD Scale) rather than a
measure of PTSD diagnosis (e.g., the SCID), resulting in estimates of PTSD
symptoms but not PTSD diagnosis. For example, Perconte (1993) assessed a

convenience sample of 620 Army, Navy, and Marine reservists. Using the MPTSD
Scale and DSM-DI-R criteria, they compared probable rates of PTSD of those
deployed (i.e., placed into battle) to the Persian Gulf to those not deployed
(Perconte, Wilson, Ponitus, Dietrick, & Spiro, 1993). The findings indicated higher
probable prevalence rates of PTSD for those deployed (15%) compared to those
who were not deployed (4%).
Sutker and colleagues (1994) used a validated diagnostic measure (DIS) and
DSM-III-R criteria to examine the relationship between extraordinary trauma
xperiences such as war-zone graves registration duty and PTSD among Operation
Desert Storm troops (Sutker, Uddo, Brailey, Vasterling, & Errera, 1994). They
found current and lifetime prevalence rates of 48% and 65%, respectively, among
troops mobilized for graves registration duty and deployed; no PTSD diagnoses
were found among the controls (demographically similar troops mobilized for
graves registration duty who did not actually serve in this capacity). This study
demonstrated the relationship between extraordinary stressors and PTSD, when
controlling for demographic factors and military characteristics.
More recently, Stretch (1996) studied a convenience sample of 4,251 active
duty and reserve veterans who were or were not deployed to the Persian Gulf War
(Stretch et al., 1996). Based on the Brief Symptom Inventory (Derogatis, 1977) and

the Impact of Events Scale (Schwarzwald, Solomon, Weisenberg, & Mikulincer,
1987), PTSD prevalence was higher for those deployed than not deployed (8%
compared to 1%). The authors suggested several possible reasons for these
differential rates when comparing deployed with those not deployed: the relative
lack of psychological preparation for deployment; and the lack of effective
debriefing and education about normal reaction to trauma exposure.
Vietnam War. Arguably, Vietnam veterans have received the most attention
in the PTSD literature. Initial estimates of the prevalence of PTSD among Vietnam
veterans were derived from the National Vietnam Veterans Readjustment Study
(NWRS) (Kulka et al., 1990), the Centers for Disease Control (CDC) Vietnam
Experience Study (The Centers for Disease Control, 1988), a Department of
Veterans Affairs Twin Study (Goldberg et al., 1990), the St. Louis site of the
NIMH-ECA study (Helzer et al., 1987), and an American Legion study (Snow,
Stellman, Stellman, & Sommer, 1988). The Department of Veterans Affairs Twin
Study, the St. Louis site of the NIMH-ECA study, and the American Legion study
used small or non-representative samples of Vietnam veterans and therefore were
not generalizable to the population of Vietnam veterans (Goldberg et al., 1990;
Helzer et al., 1987; Snow et al., 1988). A review of the remaining studies follows. 3
3 Grave registration duties involve handling, processing, and transporting human remains in support of military combat
operations and include the recovery of bodies and body parts, matching body parts to bodies, identification of remains, and
transporting bodies from the scene.

The CDC Vietnam Experience Study researchers (The Centers for Disease
Control, 1988) interviewed a large random sample of Army enlisted men who had
served one tour of duty in Vietnam. Using the DIS as the basis for identifying cases
of PTSD, the CDC study team estimated a lifetime prevalence of combat-related
PTSD of 14.7% and a current prevalence of 2.2%. In this study, they only looked
at combat-related PTSD, and they did not examine the prevalence of PTSD among
non-Vietnam veterans or civilians.
The Congressionally mandated NWRS study was the first attempt to
assess the prevalence of PTSD among a national probability sample of Vietnam
veterans and to describe comprehensively the total life adjustment of Vietnam
Theater veterans (Kulka et al., 1990). The NWRS reported six-month prevalence
rates4 of PTSD at 14.1% among male Theater veterans and 7.3% among female
Theater veterans, based on the SCID (Kulka et al., 1990). The six-month
prevalence was higher among minority veterans 22.9% among Hispanic male
Theater veterans and 19.2% among Black male Theater veterans.5
The AIWP, conducted by the National Center for American Indian and
Alaska Native Mental Health Research (NCAIANMHR), was the first
comprehensive survey to determine the prevalence rates of PTSD in American
Indian communities. Results from the AIWP, which used the SCID and DSM-III-
4 Six-month prevalence rates refer to prevalence within the last six months retrospectively from the day of the interview.

R criteria, indicated two important findings. First, current prevalence rates of
PTSD (22.1% for the Southwest group and 25.4% for the Northern Plains group)
were significantly higher than rates found by the NWRS among White veterans
(10.0%). Second, despite high current prevalence rates of PTSD among American
Indian veterans, many veterans did not develop PTSD. Even among those who had
been diagnosed with PTSD at one time, a significant portion no longer met current
criteria for PTSD. These observations suggested that other contributing factors
important to American Indian populations may affect vulnerability to PTSD
In summary, the prevalence rates of PTSD vary according to assessment,
methodology, population studied, and definition of trauma (Breslau, Davis,
Andreski, & Peterson, 1991; Kessler et al., 1995; Resnick et al., 1993); however,
some conclusions can be made. For instance, across these studies, PTSD prevalence
rates vary depending on the type, severity, and duration of the stressor. Thus, PTSD
is not an inevitable consequence of trauma exposure; rather the prevalence rates
may be related to many issues such as the nature and intensity of the stressor and
predisposing risk factors in the populations studied.
Some evidence suggests minority Vietnam veterans were at increased risk
for PTSD. Higher prevalence rates of PTSD were found among minority groups 5
5 It is important to note that the prevalence rates cited by the NWRS were based on a composite model including multiple
measures of PTSD derived from the lay and clinical interview (Kulka, 1990).

suggesting that other social or cultural factors may explain the variability in PTSD
outcomes among Vietnam veterans. Higher prevalence rates among ethnic
minorities may also be due to different levels of stress exposure; American Indian
Vietnam veterans were in heavy combat more often than other minorities (National
Center for American Indian and Alaska Native Mental Health Research, 1997).
Additional research is needed in ethnic minorities to explore possible mediating
effects of social and cultural factors on PTSD outcomes.

Table 2.1: Summary of Combat-Related PTSD Prevalence Estimates in Community Studies
World War 11 Lifetime Current
Pacific Theater Study (Sutker et al., 1993) 36 POW survivors and 29 combat veterans 40 years after war experience Diagnostic Interview Schedule DSM-III-criteria 70% (POWs); 18% (combat veterans)
40-Year Follow-Up of POWs (Kluznick et al.,1986) 188 former POWs PTSD symptom checklist administered by clinician DSM-III-criteria 66% 47%
40-Year Follow-Up of POWs (Zeiss & Dickman, 1986) 442 former POWs Self-report PTSD symptom checklist DSM-III-criteria 55%
Korean War
Combat-related Posttraumatic Stress Disorder Symptoms in Older Men (Spiro & Schnurr, 1994) 1210 veterans of World War II and the Korean War Mississippi Scale for Combat- Related PTSD .74 % among in WWII veterans .25% in Korean conflict veterans
Vietnam War
The Matsunaga Study (Beals et al., under review) 5 samples of male Vietnam veterans from the American Indian Vietnam Veterans compared to the National Vietnam Veterans Readjustment Study Sample Structured clinical interview (SCID) DSM-III-R criteria 45.3% (Southwest) 57.2% (Northern Plains) 33.7% (Hispanic) 35.4% (Black) 19.9% (White) (Past month) 22.1% (Southwest) 25.4% (Northern Plains) 22.9% (Hispanic) 19.2% (Black) 10.0% (White)
Vietnam Experience Study (CDC, 1998) 2,490 Army enlisted men with one tour of duty in Vietnam (probability sample) Survey interview (DIS) 14.7% Past month: 2.2%

Table 2.1: Summary of Combat-Related PTSD Prevalence Estimates In Community Studies (Cont.)
National Vietnam Veterans Readjustment Study (Kulka et al, 1990a, 1990b; Jordan 35 al., 1992, Weiss et al., 1992) 1,632 male and female Vietnam veterans; 716 other veterans of same era; 668 nonveterans (community epidemiologic study using probability samples) Structured clinical interview (SCID) DSM-III-R criteria 30.9 % (males) 26.9% (females) (Past six months) 27.2% (Hispanic men); 18.2 % (Black men); 14.1% (White men);
Lebanon War
1982 Lebanon War Study (Solomon et al., 1987) 382 frontline combat soldiers who had experienced acute stress reaction; 334 demographically matched frontline combat soldiers who did not experience acute combat stress reaction PTSD Inventory 1 year after war: 16% (among soldiers who had not had an acute combat stress reaction during war); 59% (among those who had experienced a combat stress reaction)
Persian Gulf War
Reservists Study (Perconte et al., 1993) 439 activated Army, Navy, and Marine reservists deployed to Operation Desert Shield/Storm Mississippi Scale for Combat-Related PTSD 15.5%deployed 4% non-deployed

Table 2.1: Summary Of Combat-Related PTSD Prevalence Estimates In Community Studies (Cont.)
Fort Devens, MA, Study (Wolfe etal., 1993) 2,344 veterans returned from Operation Desert Shield/Storm Mississippi Scale for Combat-Related PTSD 4% (men); 9% (women)
Operation Desert Storm Graveside Registration Duty (Sutker et al., 1994) 40 troops mobilized for graves registration deployed 20 control troops mobilized for graves registration non-deployed Mississippi Scale for Combat-Related PTSD 65% 48% (deployed); 0% (controls)
Active-Duty and Reservists Study (Stretch et al., 1996) 1,524 troops deployed to Operation Desert Shield/Storm 2,727 non-deployed Inpact of Event Scale and Brief Symptom Inventory 8% (deployed); 9.3% (reservists); 1% (non-deployed)

Stress-Vulnerability Theory
This section discusses the Stress-Vulnerability Theory (Dohrenwend, 1978),
which provides the conceptual framework to examine the relationship among
personal resources, trauma, and PTSD symptomatology among American Indian
Vietnam veterans.
Stress-Vulnerability Theory provides the conceptual framework for
examining PTSD in this study (Dohrenwend & Dohrenwend, 1978). Stress-
Vulnerability Theory proposes that pre-existing social conditions and personal
disposition mediate/moderate the causal relationships between stress and
psychopathology (Bloom, 1984). This theory conceptualizes stress as a major
causal factor of psychological distress. Stressors can take many formse.g., life
events, chronic strains and traumatic eventsand can vary in duration, frequency,
and severity. In this model, stress may increase an individuals vulnerability to the
effects of subsequent stress, with other factors mediating and moderating the degree
of vulnerability and response to stress. This model implies that both the stressor and
the other mediating and moderating factors may be related to psychological
distress. According to this theory, exposure to stressors triggers illness onset in
vulnerable populations. The sources of vulnerability may vary according to unique

factors within the population or individual (e.g., economic, physical, social,
cultural, psychological) that in this case may include predisposing factors (e.g.,
socio-demographic variables), psychological factors (e.g., childhood behavioral
problems), and special factors associated with American Indian populations (e.g.,
reservation life, discrimination, forced acculturation).
In addition, Stress-Vulnerability Theory allows for the social context of the
stress-illness relationship and the exploration of direct and indirect relationships
between stress and psychopathology. Moderators and mediators such as social
support and ethnic identity may have important roles in American Indian
communities as protective factors that affect ones resilience to the negative
impacts of stress. Such factors may mediate or moderate the stress-illness
relationship, as evidenced by the people who survive stressors without
psychological problems. See Figure 2 for a diagram of the Stress-Vulnerability
Model adapted for this study, from Dohrenwend's version reported in 1978.
Dohrenwend's original Stress-Vulnerability Theory discusses stressors, social
situations, predisposing factors, and health outcomes. As presented here, each
Stress-Vulnerability Theory construct corresponds with the following
hypothesized constructs for the current study: stressors refer to military and
nonmilitary trauma; social situations correspond with the hypothesized mediators
and moderators (e.g., social support and ethnic identity); vulnerability factors
include predisposing factors, psychological factors and factors unique to American

Indian populations; and psychopathology refers to the outcome measured in this
study, PTSD symptomatology.
Figure 2.1: Stress-Vulnerability Theory (Dohrenwend, 1978; Bloom, 1984)
(e.g., social support
and ethnic identity)
(e.g., military and
non-military trauma)

(e.g., PTSD symptomatology)
Vulnerability Factors (e.g.,
predisposing factors, psychological
factors, and special factors
associated with American Indian
The figure is adapted from Dohrenwend's Stress-Vulnerability Theory (Dohrenwend, 1974)

A substantial body of research has found an association between stress and
psychological distress or disorder (Kessler, Price, & Wortman, 1985; Myers, 1984;
Thoits, 1983; Turner & Avison, 1992; Turner et al., 1995b; Vingerhoets &
Marcelissen, 1988). However, the associations in the literature are not robust and
have led several researchers to investigate possible explanations for the
discrepancies in the findings. Numerous studies have examined different aspects of
stressors to explain the complexity of the stress-illness relationship. This subsection
includes a discussion of the conceptualization of stress used in this study, with a
review of the different aspects of stressors identified as important to PTSD
Conceptualization of Stress. The measurement of stress has been widely
debated, resulting in four predominant conceptualizations: trauma, chronic strains,
life events, and hassles (Aldwin, 1994). These four types of stress represent the
most basic approaches to the operationalization of stress. Although these stress
constructs may overlap, their main distinguishing features are the duration of the
event and severity of the event. First, trauma tends to be of relatively short duration
characterized by life-threatening severity and can be divided into three types:
natural and technological disasters, war and related problems, and individual

trauma. Trauma has been associated with negative mental health outcomes and has
been shown to have independent and cumulative effects on outcomes (Turner &
Lloyd, 1995). Second, conditions lasting for long periods of time that are not
immediately life-threatening can be considered chronic strains. Chronic strains
have consistently been found to have independent and cumulative effects with life
events on mental health outcomes (Pearlin, 1989; Turner & Avison, 1992). Third,
life events may be of varying duration, but they differ from chronic strains in
having clearly defined endpoints. The general consensus is that stressful life events,
especially those perceived to be undesirable and unanticipated, are disruptive and
produce stress for the individual (Dohrenwend & Dohrenwend, 1978; Kessler,
1997; Thoits, 1983; Turner & Avison, 1992; Turner & Lloyd, 1995). Fourth, a
hassle is an event of short duration that is usually minor; however, a hassle may be
embedded in the context of an ongoing life event or chronic strain. Hassles have
been associated with psychological distress and physical symptoms (Aldwin, 1994;
Delongis, Coyne, Dakof, Folkman, & Lazarus, 1982; Lazarus, 1990; Wagner,
Compas, & Howell, 1988).
Given the variety of stress measures available, the question of what type of
stress to measure should depend on the research question and the health outcome
studied. For the purpose of this dissertation, stress was measured in terms of
traumas-e.g., war, natural disasters, and individual trauma. Some examples of the

empirical findings important for justifying the stress construct used are provided
Dose-Response Relationship. Several epidemiological studies of exposure
to trauma concluded that the risk of PTSD is proportional to the intensity of
traumatic events. Furthermore, this relationship between intensity and risk was
found across a variety of settings (i.e., natural disaster, combat, criminal
victimization) and across different methodological approaches (Foy, Carroll, &
Donahoe, 1987; Green & Berlin, 1987; Helzer et al., 1987; Hiley-Young, Blake,
Abueg, Rozynko, & Gusman, 1995; Kilpatrick et al., 1989; Kiser, Ackerman,
Brown, & Edwards, 1988; March, 1993; Pitman, Altman, & Macklin, 1989; Shore
et al., 1989; Snow et al., 1988; Zeis & Dickman, 1989).
Severity and Frequency. Among veterans, trauma has been singled out as
the most significant predictor of PTSD in comparison to other pre-military,
military, and post-military factors (Breslau & Davis, 1987; Foy et al., 1987; Green
& Berlin, 1987). These studies found that enduring effects of PTSD were
associated with actual exposure to combat stressors rather than with the threat of
combat stressors. For example, in NWRS, the Vietnam veterans who had
experienced the highest levels of exposure to traumatic events in the war zone had
the highest prevalence rates of PTSD (Kulka et al., 1990). Specifically, higher
combat exposure included actual participation in atrocities (Breslau & Davis, 1987;
Green & Berlin, 1987) or exposure to brutal human death (Yehuda et al., 1992).

Subjective Perception. One explanation for the variability of PTSD
outcomes across traumatic events is the subjective nature of the traumatic event.
Researchers have argued that the perception of the event determines whether or
not the event is traumatic (March, 1993). Large variability exists across events that
may lead to PTSD, and it is difficult to categorize those events as traumatic or not
without taking subjective perceptions into consideration. The most salient
empirically documented aspects of subjective perception involve the perception of
life threat, the potential for physical violence, the experience of extreme fear, and
the attribution of personal helplessness (Green, Lindy, & Grace, 1985b; Solomon,
Mikulincer, & Flum, 1988). For example, several studies examined the content of
a stressor, such as perceived threat to life, severe physical harm or injury, exposure
to grotesque death, and loss and/or injury of a loved one. They found that the
content of the stressor was the most important predictor of PTSD and explained
the majority of the variance in comparison to other risk factors (Beals et al., under
review; Boscarino, 1995; Fontana & Rosenheck, 1994; Hiley-Young et al., 1995).
In summary, the stress construct used in this study is based on the research
question concerning social support as a mediator/moderator of military trauma and
the nature of the PTSD symptomatology outcome. The appropriateness of trauma
as the stress construct is supported by empirical evidence that trauma severity

explained the majority of the variance for PTSD. Stressors may explain some of
the variance when predicting PTSD, but not all, leading to the investigation of other
vulnerability factors that may be related to PTSD. In the next section, factors that
may account for an individual's vulnerability to developing PTSD will be explored.
These risk factors may place the American Indian veterans at extremely high risk
for PTSD.
Vulnerability Factors
In Stress-Vulnerability Theory, the sources of vulnerability vary according
to the individual and population studied, and in this study include a variety of areas:
predisposing factors (e.g., socio-economic status, education, race/ethnicity),
psychological factors, and special aspects more common among American Indian
Vietnam veterans (e.g., reservation life, discrimination, and forced acculturation).
Predisposing Factors. Certain characteristics may predispose a person to
PTSD, For example, female gender has been found to be an important predictor of
PTSD (Breslau et al., 1991; Fontana & Rosenheck, 1997; Helzer et al., 1987;
Kessler et al., 1995; Shore et al., 1989). Given that the ATWP sample contained
only males, gender will not be examined here. Other predisposing risk factors
associated with PTSD that play an important role in American Indian populations
include socio-economic factors, education, and race/ethnicity.

Socio-Economic Factors. Low socio-economic statustypically
operationalized by measures of income, occupation, or living conditionshas been
linked with negative mental health outcomes; however, the causal direction of the
relationship is unclear. For example, the prevalence of some types of psychiatric
disorders may be higher in low SES groups because their members are exposed to
more stress and adversity; on the other hand, those with psychiatric disorders may
be more likely to drift down into or fail to rise out of lower SES groups (Bruce,
Takeuchi, & Leaf, 1991; Canino et al., 1987; Holzer et al., 1986; Kessler et al.,
1994; Myers, 1984; Robins & Regier, 1991; Stansfield & Marmot, 1992; Turner &
Lloyd, 1995). Not surprisingly, American Indian populations are at high risk for
distress. They constitute a low-income, highly stressed population and,
consequently, might be hypothesized to be more likely to suffer from extensive
mental as well as physical health problems. For example, approximately 23% of
American Indians live below the poverty line when compared to 7% Whites and
26.5% Blacks (U.S. Department of Health and Human Services, 1997).
Concurrently, high rates of psychiatric morbidity were found in American Indian
adults. In a small community-based prevalence study, lifetime prevalence rates for
any psychiatric disorder were 69.4%; for affective disorders, 28.6%; and for
alcohol-related problems, 57% (Kinzie et al., 1992). Suicide rates were 5 times
higher among American Indians and Alaska Natives than among the general United

States population (May, 1987; U.S. Department of Health and Human Services,
1997). Current depression prevalence rates have been reported to be 8% in a
Northwest Coastal village, compared to rates of 0.2% to 5.2% reported in other
U.S. populations (Shore, Kinzie, & Hampson, 1973).
Education. Low education levels have been associated with psychiatric
disorder (Robins & Regier, 1991), and, on the whole, American Indians have
markedly lower levels of education than Whites. According to the 1980 census,
69% of Whites had completed 12th grade, while only 55% of American Indians had
done so (Snipp, 1992).
Race/Ethnicity. Considerable debate has arisen in the literature over the
definition and use of the term race/ethnicity. For the purposes of this research,
"race/ethnicity" is used to differentiate ethnic groups, to highlight salient
characteristics, but not to make overgeneralizations or "ethnic glosses."6
A number of articles have documented a relationship between race/ethnicity
and health status; however, the findings are inconsistent (Sondick, Lucas, Madans,
& Smith, 2000). In particular, some studies have suggested that minority Vietnam
veterans may be more vulnerable to psychological distress (Egendorf et al., 1981;
Kulka et al., 1990; Ruef, Litz, & Schlenger, 1999). Schlenger (1992), on the other
hand, found in the NVVRS that adjusting for a number of predisposing childhood
Research dealing with ethnic minorities often uses ethnic glosses or broadly refers to a specific group without
giving any attention to the heterogeneity that exists within that group (Trimble, 1991).

and family background characteristics and examining factors by time period during
the military reduced the large racial/ethnic group differences in current PTSD
prevalence rates between White/other and Hispanic males by 57% and between
Black and Hispanic male by 67%. Using the SCED and DSM-1H-R criteria, Beals
and her colleagues (Beals et al., under review) found among five ethnically defined
samples of male veterans that the two American Indian samples reported higher
lifetime prevalence rates of PTSD than did the White sample. Moreover, after
controlling for military stress exposure, the effect of ethnicity was no longer
statistically significant. Thus, the relationship between race/ethnicity and
psychological distress among minority Vietnam veterans remains unclear and needs
to be elucidated.
Psychological Factors. Empirical evidence suggests that psychological
factors, such as family history of psychiatric disorder (Davidson et ah, 1989) and
childhood behavior problems (Yager, Laufer, & Gallops, 1984), are risk factors for
PTSD in both the general population and the veteran population. Childhood
behavior problems before the age of 15 (e.g., stealing, lying, truancy, vandalism,
running away from home, fighting, substance abuse, early sexual experience) and
individual history of psychiatric disorder (e.g., anxiety disorders, depression) were
associated with increased exposure to trauma and the risk of developing PTSD
symptoms in the general population (Helzer et al., 1987). In a clinical Vietnam
veteran population, Davidson (1989) found that 66% of veterans with PTSD had a

family history of psychiatric disorder (Davidson et al., 1989). Fontana and
Rosenheck (1993,1994) suggested the following additional pre-military variables
that might be important among Vietnam veterans: a background of academic
difficulty, an unstable family environment, a father who had been in combat,
problems with the authorities, illegal drug use, and ethnic minority status.
Evidence suggests in the veteran population, that history of childhood
trauma may be an antecedent to the development of PTSD. Among a convenience
sample of Persian Gulf veterans, pre-exposure to sexual and physical abuse and
combat-related PTSD were examined. After adjustment for psychiatric history and
combat exposure, history of physical abuse was associated with higher PTSD
symptoms. These findings suggest the importance of prior trauma and the
possibility that exposure to trauma may create additional vulnerability if someone is
exposed to subsequent trauma (Breslau et al., 1998). King et al. (1996), using the
NWRS data, recently examined the relationships among pre-war factors, war-zone
exposure, and PTSD. Their findings suggest that pre-war factors such as pre-
military trauma had a direct effect on PTSD for male Vietnam veterans, when
controlling for war-zone exposure (D. W. King, King, Foy, & Gudanowski, 1996).
Special Aspects among American Indian Vietnam Veterans. This
subsection examines aspects specific to American Indian Vietnam veterans, such as
reservation life, discrimination, and forced acculturation, that may increase then-
vulnerability to PTSD.

Reservation Life. Approximately 35,000 American Indian Vietnam
veterans currently reside on more than 300 reservations across the United States
(Department of Veterans Affairs, 1992). American Indian Vietnam veterans may
be more susceptible to PTSD because of the stressful conditions that often
accompany life on the reservations. Reservation life is commonly marked with high
rates of unemployment, poverty, limited technology (for instance, a small
percentage of the population has telephones), and limited funding for health care
services (Hammerschlag, 1982; Shkilynk, 1985).
Discrimination. In addition to their economic disparity, the American
Indians are frequently subjected to damaging instances of bigotry and
discrimination (O'Nell, 1993). The National Survey of Indian Vietnam Veterans
found evidence of discrimination against American Indian Vietnam veterans
(Holm, 1994). This study found that American Indians were more often deployed
to frontline combat situations than were soldiers from other ethnic groups.
American Indian veterans were in heavy combat more often than other minorities:
75% of the Indian respondents served in infantry, airborne, tank, and artillery units,
and 42.2% saw heavy combat Results from the ATWP support these findings
(National Center for American Indian and Alaska Native Mental Health Research,
Acculturation. Acculturation refers to the degree to which one accepts or
adheres to majority values (Berry, Kim, Minde, & Mok, 1987). Forced

acculturation and oppression are continually experienced by American Indian
populations in the United States. For example, American Indian adolescents are
extremely susceptible to high levels of stress engendered by the task of identity
development, in that they may feel "caught between two cultures" (U.S. Office of
Technology Assessment, 1990). Unique stressors from the clash of cultures may
manifest themselves in psychological and physical problems (Berry et al., 1987).
American Indian Vietnam veterans are at even higher risk for PTSD due to
exposure not only to combat stressors but also to stressors from forced acculturation
In summary, high rates of stress associated with reservation life,
discrimination, and forced acculturation are found in American Indian
communities, putting American Indian populations at high risk for mental
disorders. At the same time, the impact of these general stressors varies
Mediators and Moderators of PTSD
Two mediating/moderating factors may be important in the stress-illness
relationship among American Indian Vietnam veterans: social support and ethnic
identity. It is hypothesized that high levels of these factors may enhance one's
psychosocial resilience to stress, resulting in more positive health outcomes. This
subsection examines these factors.

Social Support. The social support literature extends across several
disciplines (psychology, anthropology, sociology, and public health), includes a
variety of conceptualizations, and measures of a number of different aspects of
social relationships. Research in the area of social support began over 20 years ago
with studies by Cobb (1976) and Cassell (1976), who emphasized the importance of
social relationships to health. Cobb (1976) focused on social support as a
moderator of life stress and emphasized the emotional value (e.g., being cared for
or loved, feeling esteemed or valued) of social support processes. Cassell (1976)
emphasized that the best way to prevent disease was through the social environment
and strengthening of one's social supports, rather than decreasing one's exposure to
stressors. The empirical work that followed focused on social support and health
outcomes, in response to a wide variety of stress-inducing conditions. In this area
of stress research, a number of review articles have documented the association
between social support and health outcomes (Cassell, 1976; Cobb, 1976; Cohen &
Syme, 1985; Gottlieb, 1983; House, Umberson, & Landis, 1988; Kessler &
McLeod, 1985; Mueller, 1980; Sarason, Shearin, Pierce, & Sarason, 1987; Turner,
1983; Uchino, Cacioppo, & Keicolt-Glaser, 1996; Wortman & Lehman, 1985).
There have been discrepant research findings, such as inconsistent effects
and moderate correlations between social support and health, but reviews indicate
that these findings may vary according to the type of social support studied and the
measures used (Berkman, 1985; Cohen & Syme, 1985; Cohen & Wills, 1985;

House et al., 1988; Kasl & Wells, 1985; Uchino et al., 1996; Wallston, Alagna,
DeVellis, & DeVellis, 1983). These discrepant findings have encouraged
researchers to further explore the important features of social support to advance
this area of research.
The next section provides a review of the pertinent literature on social
support as it relates to extreme stress and PTSD symptomatology. It includes
theoretical models of social support, the conceptualization of social support,
negative aspects of social relationships, and relevant literature on social support and
PTSD among veterans.
Theoretical Models of Social Support. The predominant theoretical models
of the relationship among social support, stress, and health outcomes include the
main effects model, the stress-buffering model, and the mediating model.
(S. Cohen & Wills, 1985; Dohrenwend & Dohrenwend, 1984; Gottlieb, 1983;
House & Kahn, 1985; Kessler et al., 1985; Sarason et al., 1987; Thoits, 1983;
Turner, 1983). The main effects model argues that social support may have a direct
effect on health and may enhance health, irrespective of stress level. Therefore, a
person embedded in a supportive network of friends and family members will be
more resistant to distress than a person who is isolated and unsupported. The
stress-buffering model proposes that social support is important to well-being only
under stressful conditions (Cohen, Gottlieb, & Underwood, 2000; Turner, 1983).
The indirect model also tests alternative mediating processes that may account for

the impact of social support in health and well-being. Statistically, the indirect
effects model first tests for an interaction or buffering effect between the stressor
and social support in predicting the health outcome, and then tests for mediating
effects if moderating effects are not found. The direct or main effects model looks
for a direct relationship between social support and the health outcome.
The literature supports both the direct and indirect effect models. These
inconsistent findings may result from different conceptualizations of social support
(e.g., structural and functional support) and types of stress measures used (e.g.,
trauma, chronic strain, stressful life events) (Cohen & Wills, 1985). Structural
support is hypothesized to have main or direct effects with health outcomes, and
functional support is hypothesized to buffer the effects of stress when predicting
health outcomes (Cohen & Wills, 1985). Other explanations suggest that there is an
optimal match between the type of social support and the stressor, such that specific
types of social support meet individual needs based on the type of stress exposure
(Cutrona & Russell, 1990).
Conceptualization of Social Support. Social support has been
conceptualized in terms of the structure of the interpersonal relationship or social
network or in terms of the function that a relationship or network serves. Structural
measures include marital status, number of relationships, number of contacts with
the family, participation in community organizations, and involvement in social
networks. The functional aspects of social support commonly fall into four areas:

1) esteem or emotional support; 2) informational support; 3) social companionship;
and 4) instrumental help or material aid (Cohen & Syme, 1985). Structural
measures are considered relatively objective and provide measures of
embeddedness in a social system; functional measures are more subjective and
describe an individual's perception of their support systems.
Functional aspects of social support are the focus of this study, and an in-
depth description of the types of functional social support is provided next.
Emotional support involves resources such as having someone to confide in or
someone to talk to about problems or personal issues. Studies of emotional support
during the military have found a large difference in symptomatology between those
who have confidant relationships and those who do not. For example, emotional
support from unit members and officers helped sustain each unit member's morale
and prevented him from feeling cut-off or overwhelmed by the extreme threats
surrounding him (Solomon, Mikulincer, & Hobfoll, 1986a). Informational support
refers to information, advice, and guidance provided by others concerning possible
solutions to a problem. For example, social networks may serve this function by
being an important source of referral information. Social companionship includes
enjoyable social activities, social visiting, dinners, parties, films, concerts, and
informal athletics. Social companionship has been shown to be important in
predicting well-being (Lewinsohn & Amenson, 1978; London, Crandall, & Seals,
1977; Spanier & Lewis, 1980). Instrumental support includes sources of support

such as providing assistance with household chores, running errands, giving a ride,
donating money, or providing material goods. Instrumental support was cited as
important to veterans because it ensured safety and organization among Israeli
soldiers amidst the extreme stress of combat (Solomon et al., 1986a).
These studies illustrate the importance of functional support in predicting
health outcomes and suggest that certain types of social support may be beneficial
in response to extreme stress. They also suggest that the relationship of social
support to health outcomes may depend on the nature of the s tressor, and the extent
to which the social support can fulfill the individual's needs at the time. The next
subsection reviews how social support can have negative effects on health
Negative Aspects of Social Relationships. The negative aspects of social
relationships and health have emerged in the literature over the last decade
(Gottlieb, 1983; Rael, Stansfeld, Shipley, & Head, 1995; Revenson, Schiaffino,
Majerovitz, & Gibofsky, 1991; Stansfeld, Rael, & Shipley, 1997; Turner & Lloyd,
1995; Turner et al., 1995b). This emphasis is mainly attributed to a
multidimensional view of social support that allows for the possibility that social
support may have both positive and negative effects on health and well-being.
Caplan (1984) described negative aspects of social relationships as "dissaffirmation
of a persons thoughts, beliefs, feelings, the impression of negative regard,
disaffection, and certain kinds of withholdings of aid or erection of barriers"

(Caplan et al., 1984). Empirical findings suggest that negative interactions
constitute an important share of the stressors in peoples' lives and, in fact, were
more strongly associated with mental health outcomes than were positive elements
of social relationships (Helgeson, 1993; Holahan, Moos, Holahan, & Brennan,
1997; Rael et al., 1995; Revenson et al., 1991; Rook, 1990; Sarason, Sarason, &
Pierce, 1990; Stansfeld et al., 1997; H. A. Turner, 1994). The most frequently cited
sources of negative social relationships are within the social network, received
support, and perceived support (Ell, 1996; Gottlieb, 1983; Helgeson, 1993; Holahan
et al., 1997; Rael et al., 1995; Revenson et al., 1991; Rook, 1990; Sarason et al.,
1990; Stansfeld et al., 1997).
In some studies, the family members (e.g., caregivers) who provide support
may be a source of stress for recipients (e.g., ill or elderly family members)
(Browne, 1982; Dunkell-Schetter & Wortman, 1981; Fisher & Nadler, 1982;
Stephens, Norris, Kinney, Ritchie, & Grotz, 1988). For example, Dunkel-Schetter
and Wortman (1981) suggested that the elderly often fail to receive support because
their suffering gives rise to the feelings of vulnerability and helplessness among the
other family members who are trying to help. In regard to ill family members, the
family may fail to come to another family members aid because of conflicting
feelings about illness and the draining of the caregivers resources. For example, in
a breast cancer study, only 50% of the breast cancer patients found their family
members to be supportive. Instead the patients felt avoided, feared, and

misunderstoodsupport did not materialize at all, or it met the emotional needs of
the other family members, not the victim or the ill family member (Peters-Golden,
In other circumstances, family members (support providers) may intend for
their interactions with ill family members to be supportive; however, the ill family
member may perceive these interactions as unnecessary or unwanted. The act of
providing support may be loaded with unwanted obligations and feelings of
dependence. These findings suggest that rendering support may be stressful both
for the family member (support provider) and the ill family member (support
recipient). Therefore, examining the negative effects of social relationships may
provide a fuller picture of the kinds of social relationships that are likely to
influence social support and contribute to a model for PTSD symptomatology.
Social Support and Veterans. Social support in Vietnam veterans has been
defined in many ways: marital status (Egendorf et al., 1981), family environment
(Stretch, 1985), unit-cohesion (Fontana, Rosenheck, Robert, & Horvath, 1997),
social network (Green & Berlin, 1987), instrumental and emotional support
(Solomon et al., 1986a), family support (Solomon & Benbenishty, 1986b), social
isolation (Kulka et al., 1990), homecoming reception (Keane, Scott, Chavoya,
Lamparski, & Fairbank, 1985; Solomon, Mikulincer, & Hobfoll, 1987c), social
companionship (Boscarino, 1995; Egendorf et al., 1981; Keane et al., 1985), and
structural and functional support (King et al., 1998). The following review of the

social support and veteran literature is organized by the military time-frame (e.g.,
military, post-military, and at the time of the interview) rather than the type of
social support.
Social Support Purine the Military. Support during the military may be an
important predictor of PTSD. Card (1983) proposed that the lack of social support
during the military was due to the length of the tours of duty in Vietnam. The tours
were shortone year of intense combatand, therefore, inhibited the development
of bonds with other soldiers (Card, 1983). A lack of social support during the
military was also related to intense combat-related PTSD (Barrett & Mizes, 1988;
Fontana et al., 1997); however, this study did not examine an interaction of combat
stressors and social support predicting PTSD. More recently, Fontana (1997)
reported a significant interaction effect between unit cohesion and military trauma.
Fontana (1997) found that under conditions of high levels of stress, high levels of
unit cohesion were associated with higher PTSD symptoms (Fontana et al., 1997).
These results suggest the possible negative effects of social support during the
military under conditions of high stress.
Post-Military Social Support. Post-military social support has been the main
focus of most of the studies that examine the relationship between social support
and PTSD among Vietnam veterans. Post-military social support has
predominantly been measured in terms of social networks, perceived support at
homecoming, family support, and structural and functional support. The

composition of the social network was found to be an important component of
perceived support. Egendorf et al. (1981) found veterans who had access to social
networks comprised of veterans were better adjusted than were veterans who had
access to social networks consisting of non-veterans. They concluded that, for
veterans, empathy and understanding were important components of emotional
social support; and that networking with other Vietnam veterans may buffer the
stress of past war experiences (Egendorf et al., 1981). This study emphasizes the
important mediating effects of emotional social support for Vietnam veterans
exposed to extreme stress.
Several studies citing negative aspects of social support, reported an
association between decreased social support at homecoming (e.g., perceived
helpfulness of veterans family at homecoming) and PTSD among Vietnam veterans
(Barrett & Mizes, 1988; Frye & Stockton, 1981; Green & Berlin, 1987; Keane &
Fairbank, 1983; Stretch, 1985). For many American Indian Vietnam veterans, in
particular, the homecoming reception was a time of emotional detachment, pain,
and confusion. Holm (1992) explained that American Indian veterans who returned
home from the war expected to find greater acceptance into mainstream society
because of their dedication and contributions to the war efforts (Holm, 1992).
However, at homecoming, "if they sought acceptance by whites, they were

disappointed. If they thought military service would bring them opportunity, they
discovered that it had only lowered their status within American society" (Holm,
1992, p. 31). Consequently, the homecoming experience for American Indian
Vietnam veterans may have placed them at increased risk for PTSD, and negative
social support during homecoming may play a significant role in prediction of
PTSD among this population.
King et al (1998) tested the stress-buffering hypothesis using the NWRS data
and examined the relationship among functional and structural social support, war-
zone exposure, stressful life events, and PTSD symptoms. They found that for both
functional and structural social support strongly mediated stressful life events, but
support for moderating effects were not indicated (D. W. King et al., 1996).
Current Social Support. Boscarino (1995) reported that Vietnam veterans
with lower levels of current social support and emotional support (e.g., number of
current close friends, satisfaction with relationships) were at greater risk for current
PTSD (Boscarino, 1995). Vietnam veterans with low social support had nearly an
80% greater risk of PTSD than veterans with average social support and had nearly
a 180% greater risk than veterans with high social support. In addition, supportive
marital relationships have been associated with successful adjustment to delayed
stress reactions from combat exposure. Men who were exposed to high levels of
combat were at increased risk for divorce and were generally less happy and less
satisfied with their lives, their marriages, and their role as fathers (Stellman,

Stellman, & Sommer, 1988). Egendorf et al. (1981) also found differences in
adjustment between married and non-married veterans, concluding that married
veterans fared better than non-married veterans, although married veterans with
little spousal support fared worse than men not married at all.
In summary, the literature on social support and Vietnam veterans indicates
that social support and PTSD are related among Vietnam veterans. Military, post-
military (e.g., perceived social support at homecoming, social networks comprised
of veterans) and current social support (e.g., supportive marital relationships, family
environment, number of current close friends, current emotional social support,
instrumental social support, and social companionship) were significant variables in
predicting PTSD outcomes. Post-military social support factors, in particular,
played an important role in the prediction of PTSD. These findings indicate that,
after taking into account the war experience, other factors may also affect the risk
of post-war problems. Furthermore, the unique characteristics of the Vietnam War-
-the political climate, the short tours of duty, and the lack of social support at
homecomingmade it difficult for Vietnam veterans to reintegrate into society.
Although the evidence suggests social support and PTSD are related, specific
relationships among social support, trauma, and PTSD symptomatology have not
been determined.

Ethnic Identity. Ethnic identity may be an important moderator/mediator
for adjustment on many levels. Consequently, it is important to discern the role that
ethnic identity may have in relationship to psychological distress. It is
hypothesized that ethnic identity may mediate psychological distress in American
Indian populations, which are at greater risk because of reservation life conditions,
discrimination, and forced acculturation. The following subsections will describe
areas important to ethnic identity: the definitions of ethnic identity, the conceptual
framework of ethnic identity, and a brief review of the literature of ethnic identity
and PTSD.
Definitions of Ethnic Identity. In a review of 70 articles on ethnic identity,
Phinney (1990) emphasized the fragmentary and inconclusive nature of the research
in this area. Most of the empirical work suffers from the following problems: no
widely agreed-upon definition of ethnic identity, very few reliable and valid
measures of ethnic identity, and ethnic identity constructs that cannot be compared
across groups. Ethnic identity has been defined in the general population in several
ways: 1) self-identification or self-labeling (Phinney, 1990); 2) feeling of belonging
and commitment to an ethnic group (Parham & Helms, 1985); 3) a sense of shared
values and attitudes with an ethnic group (Parham & Helms, 1985); 4) cultural
aspects of ethnic groups such as language, behavior, values, and knowledge of
ethnic history (Caltabiano, 1984; Leclezio, Louw-Potgieter, & Souchon, 1986;

Phinney, 1990); and 5) a persons affiliation with an ethnic group (Moran, Fleming,
Somervell, & Manson, 1999; Oetting & Beauvais, 1991).
Conceptualization of Ethnic Identity. Two models provide the conceptual
framework for ethnic identity. One model proposes that a persons ethnic identity
falls along a single continuum from identification and involvement with an ethnic
group to identification with the mainstream (Bumam, Telles, Kamo, Hough, &
Escobar, 1987; Phinney, 1990). This view requires that a person fall somewhere on
a unidimensional continuum in which the presence of one type of ethnic identity
presumes the absence of another. The second model addresses a multi-cultural
view of ethnic identity, where people may be capable of identifying independently
with more than one ethnic group at the same time (Oetting & Beauvais, 1990-
1991). This modelthe orthogonal cultural identity modelencourages
measurement of concurrent identification with more than one reference group,
allowing one to view ethnic identity as a dynamic process.
Ethnic Identity and PTSD. Beiser and Turner (1989) examined the effect of
ethnic identity on PTSD outcomes in Southeast Asian refugees. Ethnic identity was
defined as marriage within the ethnic group and social support provided by it. They
concluded that marriage may moderate the risk of developing depressive symptoms
associated with PTSD by enhancing a sense of "self-identity" and belonging (Beiser
& Turner, 1989). In a study of Guatemala Mayan children who were trauma
survivors, Melville (1992) identified ethnic identity factors that helped the Mayan

children survive severe trauma (Melville & Lykes, 1992). Some socio-cultural
characteristics identified included ethnicity of social network (family and friends),
and community-specific language. In summary, the literature suggests that the
degree to which one identifies and participates in traditional cultural lifestyles may
relate importantly to PTSD outcomes.
Empirical findings on ethnic identity and PTSD in the veterans literature
are sparse. Definitions of ethnic identity have included negative attitudes and
alienation from one's ethnic group and cultural traditions. In a study of Hispanic
Vietnam veterans, Escobar and colleagues (1983) reported "highly symptomatic
PTSD veterans reported significantly smaller networks, fewer contacts outside the
family circle ... PTSD veterans appeared more alienated from their cultural
heritage than the other groups" (Escobar, 1983, p. 585). Thus, for Vietnam
veterans from ethnic groups, high levels of ethnic identity may be a protective
resource that mitigates stress (Escobar et al., 1983).
Other studies examined cultural traditions in two samples, Japanese-
American Vietnam veterans, and American Indian Vietnam veterans. Hamada and
colleagues (1988) described the unique difficulties associated with being a
Japanese-American in Vietnam. Cultural traditions regarding the presentation of
one's self that distinguishes one's true feelings ("honne") from one's public
expression of feeling ("tatemae") may lead to cross-cultural miscommunication and
misunderstanding. They concluded that the unique difficulties associated with

being misunderstood resulted in low levels of ethnic identity, which increased a
veterans vulnerability to PTSD (Hamada, Chemtob, Sautner, & Sato, 1988).
Studies examining ethnic identity among American Indian veterans have also
assessed ethnic identity in terms of cultural traditions. Cultural traditions such as
the "warrior concept," especially among Plains American Indians, may limit
motivation to seek assistance for personal problems because of an emphasis on
stoicism, self-denial, and personal pride O'Nell, 1994a). Thus, American Indian
veterans with high levels of ethnic identity may have increased vulnerability to
PTSD by interfering with help-seeking behavior.
In summary, the literature on ethnic identity suggests a potential
relationship between ethnic identity and PTSD among the American Indian
population and recommended further exploration of the ethnic identity construct for
this population. The present study used the AIWP items adapted from the
questionnaire measuring ethnic identity developed by Oetting and Beauvais (1990-
1991) and adapted by Moran et al. (1999). This scale includes questions about
ethnic beliefs, activities, and language.
The literature on social support and ethnic identity suggests future work in
this area can improve our understanding of trauma, social support, ethnic identity,
and PTSD symptomatology among American Indian Vietnam veterans. Several
areas have been identified as important areas for future research in this population.
The specific aims of this study addresses the following areas:

1. To explore the relationships between pre-military predisposing factors and
PTSD symptomatology.
2. To explore the relationships among military characteristics, social support
during the military, severity of military trauma, and PTSD symptomatology.
3. To examine the relationships between post-military social support and
PTSD symptomatology.
4. To explore the relationships between moderators and mediators of military
trauma and PTSD symptomatology.
5. To explore the relationships among current social support, current ethnic
identity, current demographics, and PTSD symptomatology.
6. To explore the relationships among current demographics and PTSD
7. To test the full model across all specific aims, and examine the existing
significant relationships determined in the previous aims.

This chapter describes the general framework for the present study,
beginning with an overview of the American Indian Vietnam Veterans Project
(AIWP). The research design, study population, sampling strategy, and data
collection procedures of the project are summarized. All AIWP measures were
adapted or modified during extensive focus group review to ensure the cultural
appropriateness of all items. This chapter describes the scale development
procedures and the psychometric properties of the final measures, to demonstrate
the reliability and cultural validity of all measures.
General Framework of the Present Study
Secondary data analyses of the AIWP dataset tested a set of hypotheses
concerning the complex relationships among trauma, psychopathology, and
hypothesized mediators and moderators of these two constructs. Stress-
Vulnerability Theory (SVT; Bloom, 1984; Dohrenwend & Dohrenwend, 1978);
which describes interrelationships among trauma, moderators/mediators, and
predisposing factors; provided the theoretical foundation for this study and guided
all subsequent analyses. Various multivariate procedures were employed, including
correlations, exploratory factor analyses, reliability analyses, and hierarchical linear

regressions. An overview of the ATVYP is discussed next as background for the
current study.
ATVYP Research Design
Community confidentiality is as important as individual confidentiality in
American Indian groups; therefore, general cultural descriptors were used rather
than specific tribal names. Two distinct cultural groups will be referred to
throughout this study, the Northern Plains and the Southwest. The Northern Plains
consists of three highly related tribes and the Southwest consists of one tribe. For
the purposes of this study, each cultural group will be referred to as a distinct tribe.
The ATVVP comprised four phases. Phase I involved development^
registries of American Indian Vietnam Theater veterans, from both the Northern
Plains and the Southwest, who were still living on or near their reservation. Phase
D used qualitative methodology to ensure the cultural validity of the instruments
and maintain comparability with the National Vietnam Veterans Readjustment
Study (NVVRS). Several focus group sessions generated a number of changes to
the instruments: culturally understandable directions, items, and prompts; culturally
appropriate interpretations of items contained in the instrument; and removal of
culturally inappropriate items. In Phase III, the adapted NVVRS instrument and the
University of Michigan version of the Composite International Diagnostic
Instrument (UM-CIDI) were administered by lay or non-clinically trained

interviewers to study participants from the Northern Plains and the Southwest.
Data collection commenced in April 1993 (Northern Plains) and in December 1993
(Southwest); all data collection was completed in May 1995. Phase IV consisted of
data analyses and preparation of the Final Report (NCALANMHR, 1997) for
dissemination. Follow-up to the communities that participated in the AIWP was
an integral component of the research study. Community education materials
summarizing the results of the study were prepared in conjunction with on-site
workshops conducted in several locations throughout the participating
communities, with the purpose of increasing awareness of PTSD among American
Indian Vietnam veterans.
AIWP Study Population
Research Sites. The research sites for the AIWP data collection consisted
of four reservations located in the Southwest and the Northern Plains. The
reservations are in isolated rural areas, with high unemployment, limited access to
technology, and limited resources for mental health services; residents of the
reservations have poor health (National Institute of Mental Health National
Advisory Mental Health Council, 1991; U.S. Department of Health and Human
Services, 1997). In addition to the economic stressors placed on their communities,
American Indians have been subjected to years of cultural oppression (Project,

Sampling Frame. The sampling frame for the AIVVP consisted of male
Vietnam Theater veterans from the targeted communities who served on active duty
in the U.S. Armed Forces between August 5, 1964, and May 7,1975. Female
Vietnam Theater veterans were excluded because of low numbers veterans within
these populations. Other eligibility criteria for the AIWP included tribal
enrollment in the Northern Plains or Southwest tribes, year of birth between 1930
and 1958, and residence within 100 miles of the reservation. The sampling strategy
for each research site is discussed below.
AIWP Sampling Strategy
Northern Plains. Sampling procedures included an extensive location and
tracking phase to identify the sample and to verify the veteran status of all men who
met other eligibility criteria. Three data sources were used during the location
phase to facilitate the identification process: tribal enrollment records, state bonus
records, and local veterans lists. First, the tribal enrollment records, maintained by
the Bureau of Indian Affairs (BIA), identified all recognized members of a tribe.
To be included on the tribal enrollment records, a person must demonstrate that he
or she is related to someone listed on the original census of that tribe and that he or

7 8
she meets the minimum blood quantum level. Second, state bonus records
provided a comprehensive list of Vietnam Theater and Vietnam Era veterans in the
state in which the Northern Plains reservations are located. Third, local veterans in
two of the three different tribes on the Northern Plains reservation developed lists
of fellow Vietnam veterans. These lists were used to identify, from the tribal
enrollment records, men who were Vietnam veterans. Because of the high profile
of the veteran in the American Indian communities, veteran status was common
knowledge. Thus, these local veteran lists provided a valuable source for
identifying the Vietnam veterans. In addition, every participant provided a copy of
his military discharge papers (DD-214) to confirm Vietnam veteran status. The
final sample frame was made up of two lists: tribal enrollment records that matched
by name with either the state bonus records or the local veterans lists (N=605), and
tribal enrollment records that did not match to any of the lists (N=719)
Sample size estimates from key-informant interviews and census data
determined that between 200 and 350 male Vietnam Theater veterans lived on or
near the three NP reservations. Thus, 100% of the sample or a population sample
was needed to ensure adequate power for prevalence estimates. Because of the
small population size, it was necessary to sample from both the matched and non- 7
7 Blood quantum level refers to degree of American Indian ancestry, with 25% American Indian blood the most commonly
accepted minimum threshold for tribal membership (T. P. Wilson, 1992). American Indians are the only ethnic group in
the United States legally defined by degree of ancestry.
* A statewide program for Vietnam Theater and Era veterans provided a monetary bonus upon separation from the military.

matched lists to ensure representation of all Vietnam Theater veterans currently
living on or near the three reservations.
During the pilot location process, a total of 362 men met eligibility criteria
from the matched (N=242) and non-matched lists (N=120). Altogether, 93%
(N=1232) men were located from the matched and non-matched lists (N=1324). By
interviewing those from the non-matched list (N=109) and from the matched list
(N=196), a total of 305 men were interviewed. Table 3.1 presents the pilot location
results from the Northern Plains.
Table 3.1: Northern Plains Pilot Location Results
Northern Plains Non-matched N=605 Matched N=719
Number located 540 692
Number eligible 120 242
Number ineligible 420 450
Number interviewed 109 196
Southwest. The data sources used for the tribe on the Southwest reservation
included tribal enrollment records and listings from the Department of Southwest
Veterans Affairs (DSVA). Two lists made up the final sampling frame: tribal
enrollment records of those included on the tribal DSVA lists, and tribal enrollment
records those not included on the tribal DSVA lists. Men from the matched list
(N=l,330) and non-matched list (N=14,072) were divided into 10 replicates based
on the last digit of their tribal identification number: Each replicate contained 10%
of the sampling frame. To begin the pilot location, eight replicates of 10% each

(80%) of the matched sample were selected (N=l,064); two replicates of 10% each
(20%) were selected (N=2,802) of the non-matched sample. During the pilot
location process, the rate of eligible veterans from the non-matched list was 1.1%.
Therefore, for every 100 men, only 1.1 met our eligibility criteria. It was cost-
prohibitive to continue to locate eligible veterans from the non-matched list at this
low rate of eligibility. Consequently, the non-matched sample was not included in
the prevalence estimates. Table 3.2 provides the pilot location results from the
Table 3.2: Southwest Pilot Location Results
Southwest Non-matched N=2,802 Matched N= 1,064
Number located 1,085 951
Number eligible 12 355
Number ineligible (including cannot 1,073 596
locate) Number interviewed 9 316

AIWP Data Collection Procedures
Location and Lay-Interview Procedures. Field workers known as
"locators" were assigned the task of locating and ascertaining veteran status of
potential respondents. If a particular veteran was found to be eligible, an interview
was scheduled. The locator kept records of all contact attempts, identified eligible
respondents, and scheduled the AIWP interview. The location rates for the
veterans were 93% for the Northern Plains sites and 89% for the Southwest site.
The interview rates were slightly lower: 84% and 89% respectively. Table 1.3
provides a summary of the ATWP location and interview efforts (National Center
for American Indian and Alaska Native Mental Health Research, 1997).
Table 3.3: AIWP Location and Lav-Interview Results
Northern Plains Southwest
Number of locators hired/ trained 8 17
Average number of locations per person 222 193
Average cost/location $47.20 $86.76
Number of interviewers hired/trained 6 14
Average number of interviews per person 66 40
Average cost/interview $130.20 $178.70
Total selected 1324 1064
Total located 1232 951
Location rate 93% 89%
Number eligible 362 355
Interviews completed 305 316
Interview rate 84% 89%

Training and Quality-Control. Lay-interviewers received two weeks of
training, immediately prior to the initiation of data collection at a given site. The
main objective of the training was to ensure the interviewer's ability to administer
all instruments with a 95% accuracy rate. Training included group practice sessions
with American Indian respondents, reviews of videotaped interviews, role-playing
with mock interviews, didactic coverage of the paper instrument, and supervised
practice interviews. An Interviewer Training Manual was developed as an
important tool for training and later reference. Periodic "booster" training sessions
were conducted in the field as quality control procedures identified areas of
difficulty with the protocol.
At the beginning of data collection, when the potential for error was highest,
NCAIANMHR staff reviewed the first three audiotaped interviews for each
interviewer to assess data quality. Early detection of errors and further instruction
prevented the repetition of errors. Interviewers were required to meet the following
criteria: 95% accuracy in reading the questions as written, 95% accuracy in
recording the respondents answers, and 95% accuracy in following the interviewer
instructions. Re-interviews were required if these standards of accuracy were not

The AIWP and NVYRS retrospectively measured key constructs defined
in Stress-Vulnerability Theory (SVT): predisposing factors, stress,
moderators/mediators (social support, ethnic identity), and outcomes (PTSD
symptomatology). Several time frames were assessed: pre-military, military, post-
military, and current status. For this study, these constructs were analyzed by time
frame to facilitate comparison with the existing literature.
This section describes the measures used in this study and the procedures
used to develop scales from the raw data. Coding, the construction of variables,
and psychometric properties of the final measures selected for all analyses are
included. A theoretically informed empirical approach was used to refine the
existing measures and to ensure the reliability and the cultural validity of all
Scale Development Procedures
A three-stage process was used to develop all scales in the present study.
First, each scale was subjected to extensive item-level psychometric analyses to
identify salient items for scale construction. Item-mean comparisons between tribes
identified any items that may have behaved differently across groups and detected

any differences between tribes. Item correlations between tribes assessed
consistency across tribes and identified general relationships between items.
Second, exploratory factor analyses with a principal components analysis followed
by a varimax rotation, determined the factor structures of all scales. Kaiser (1960)
criterion and Catell (1966) guidelines for the scree test suggested the number of
meaningful factors to be retained for rotation (Cattell, 1966; Kaiser, 1960). In
interpreting the factor pattern, an item was said to load on a given factor if the
factor loading was .40 or greater. All items were then tested for internal
consistency as indicated by Cronbach's alphas (Devellis, 1991). After several
reiterations of psychometric and factor analyses, final sets of items were selected.
Third, second-order principal components analyses were warranted if at least three
components were indicated from the initial principal component analyses.9
Collapsing subscales into one multidimensional scale was recommended if subscale
correlations were moderate to high and the second-order components analyses
indicated a one-component solution. For scales with the same response sets, an
average-item, unit-weighted scale score was created for all items, adjusting the
items for the number of non-missing observations. For the scales with items that
had different response sets, items were standardized as described by Spector (1992).
Z-scores were calculated for each item, summed and adjusted for the number of
non-missing items (Spector, 1992).
SAS requires a minimum of three items or components to conduct exploratory factor analyses (SAS Institute, 1990).

Each of the measures used is discussed below, organized by SVT
construct (e.g., outcome variable, predisposing factors, social support, ethnic
identity, and stress) and time frame. The scale construction sections of PTSD
symptomatology, social support, and ethnic identity include item means, item
correlations by tribe, factor analyses, and reliabilities. First, item characteristics
are provided which include a complete list of scale items, response set, number of
non-missing responses, and item means by tribes. Second, item correlations by
tribe are presented to determine the relationship between items, and to check for
consistency between tribes. Third, principal components analyses are presented
suggesting factor structures of the items. If necessary, second-order factor
analyses follows with a recommendation to collapse the final scale. Cronbach's
alphas are also presented to test for internal consistency. Items listed in the tables
(e.g. SRK29a, SRK29b) refer to the corresponding item number in the AIWP
interview. Final scale variable names are capitalized (e.g., ED ATTAIN). All
final measures are summarized at the end of each section. (See Appendix 2 for
specific items and actual scales from the AIWP lay-interview.)

Outcome Variable
PTSD symptomatology, assessed with the Mississippi Combat-related
PTSD (MPTSD) scale, was the outcome (dependent) variable in all analyses. The
MPTSD scale is a continuous measure of symptom severity, selected because of its
demonstrated reliability and strong associations with the clinical diagnosis of PTSD
(Keane & Penk, 1988; Kulka et al., 1990). This section includes a discussion of the
distinction between PTSD symptomatology and PTSD diagnosis, justification for
using PTSD symptomatology as an outcome measure in this study, characteristics
of the MPTSD scale, and construction of the MPTSD Scale.
PTSD Svmptomatology/PTSD Diagnosis. It is important in traumatic stress
research to distinguish between PTSD diagnosis and PTSD symptomatology.
PTSD diagnosis, a dichotomous measure, refers to the presence or absence of
PTSD a distinction based on DSM diagnostic criteria. PTSD diagnostic status is
useful in several situations: in epidemiology studies, to determine current and
lifetime prevalence rates of a disorder; in case-control studies, to compare
characteristics between cases and non-cases; and in treatment-outcomes studies, to
assess diagnostic status and effectiveness of treatment models. PTSD
symptomatology, on the other hand, provides a continuous measure of symptom
severity that is useful for research studies that, like this one, address the relationship

between PTSD symptomatology and other ordinal and continuous measures such as
trauma and personal resources. As a continuous measure, it is also useful in
detecting subtle changes in symptomatology and associated features of PTSD. The
MPTSD Scale used in this study includes items assessing associated features of
combat-related PTSD, including depression, substance abuse, and impairment of
social functioning.
The heterogeneous nature of PTSD recommends the use of PTSD
symptomatology rather than diagnostic criteria in some research studies. The onset
of PTSD symptoms may be immediate or delayed, and the longitudinal course may
be acute, chronic, recurrent (intermittent), reactivated or residual (Blank, 1993;
Bremner et al., 1993; Green et al., 1990). Some individuals can have sub-threshold
or partial PTSD with substantial impairment without meeting the full criteria for the
PTSD (Kessler et al., 1995; Resnick et al., 1993). In addition, researchers have
described a complex post-traumatic syndrome in survivors of severe, prolonged
trauma (American Psychiatric Association, 1994; Herman, 1992; Pelcovitz, 1997;
Roth, Newman, Pelcovitz, van der Kolk, & Mandel, 1997). These studies suggest
that current PTSD diagnostic criteria may not accurately represent the full range of
symptoms of PTSD, emphasizing some advantages of assessing PTSD
symptomatology over diagnostic criteria.
Characteristics of the MPTSD Scale. The MPTSD assesses current
symptoms of PTSD as they are delineated in DSM-ID-R (Keane et al., 1988). The

MPTSD Scale consists of 35 items, scored on a Likert-type scale, and recoded so
that higher scores equaled greater severity. Due to the small percent of missing
cases (less than 1%), missing data were handled with listwise deletion. Table 3.4
provides a complete list of the scale items, response set, number of non-missing
responses, and item means across tribes. This is followed by the psychometric
analyses of these items, including an overview of the item means, item correlations,
principal component analyses, and internal consistencies used to construct the final
Item Means. Comparisons between means indicated that the majority of the
items were not significantly different across tribes. The results did suggest,
however, a few significant differences across tribes. For example, the Southwest
group was less likely to report being even-tempered and easy-going (SRK29aa),
and more likely to report being upset by reminders of military service (SRK29d).
In addition, the Southwest veterans were more likely than the Northern Plains to
endorse some re-experiencing and situational avoidance items. For instance, they
endorsed an inability to remember one's military experience (SRK29kk) and the
perception that no one understands feelings (SRK29z). On the other hand, the
Northern Plains group were more likely to report wondering why they were still
alive (SRK291), jumping at unexpected noises (SRK29y), and having memory
rarely as good as before (SRK29hh).

Table 3.4: MPTSD Scale Item Characteristics
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Before service had more friends (SRK29a) 1 = Not at all true 620 Total Sample 2.48 (.05)
2 Slightly true 315 Southwest 2.43 (.07)
3 - Somewhat true 4 = Very true 5 = Extremely true 305 Northern Plains 2.55 (.08)
R is even-tempered and easy-going (SRK29aa) 1 = Never 1 = Always 618 Total Sample 2.39 (.04)
2 = Rarely 2 = Usually 313 Southwest 2.53 (.06)
3 = Sometimes 3 = Sometimes 305 Northern Plains 2.25 (.05) ***
4 = Usually 4 = Rarely
5 = Always 5 Never
No guilt over things did service (SRK29b) 1 = Never true 1 = Always true 620 Total Sample 3.11 (.05)
2 = Rarely true 2 Usually true 315 Southwest 3.06 (.07)
3 = Sometimes true 3 = Sometimes 305 Northern Plains 3.17 (.07)
4 = Usually true true
5 = Always true 4 = Rarely true 5 = Never true
No one understands military experience (SRK29bb) 1 Not at all true 618 Total Sample 2.99 (.05)
2 = Slightly true 314 Southwest 3.06 (.08)
3 = Somewhat true 4 = True 5 = Very true 304 Northern Plains 2.91 (.08)
R likely to become violent (SRK29c) 1 = Very unlikely 620 Total Sample 3.38 (.04)
2 = Unlikely 315 Southwest 3.45 (.06)
3 = Somewhat unlikely 305 Northern Plains 3.31 (.07)
4 = Very likely 5 = Extremely likely
Use drugs to forget military experience (SRK29cc) 1 = Never 619 Total Sample 3.02 (.06)
2 = Infrequently 315 Southwest 2.99 (.07)
3 = Sometimes 4 = Frequently 5 = Very frequently 304 Northern Plains 3.04 (.07)
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE = standard error
*p <. 05, ** p<.01, *** p<001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics (Cont.l
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Reminders of service upset (SRK29d) 1 = Never 619 Total Sample 2.92 (.04)
2 = Infrequently 314 Southwest 3.01 (.06)
3 = Sometimes 305 Northern 2.82 (.06) *
4 = Frequently Plains
5 = Very frequently
Comfortable when in a crowd (SRK29dd) 1 = Never 1 = Almost always 619 Total Sample 3.38 (.05)
2 = Rarely 2 = Usually 315 Southwest 3.44 (.06)
3 = Sometimes 3 = Sometimes 304 Northern 3.32 (.07)
4 = Usually 4 = Rarely Plains
5 = Almost always 5 = Never
People who know R fear him (SRK29e) 1 = Never true 620 Total Sample 2.25 (.05)
2 = Rarely true 315 Southwest 2.21 (.06)
3 = Sometimes true 305 Northern 2.29 (.06)
4 = Frequently true 5 = Very frequently true Plains
Explode over minor things (SRK29ee) 1 = Never 618 Total Sample 2.47 (.04)
2 = Infrequently 313 Southwest 2.52 (.06)
3 = Sometimes 305 Northern 2.42 (.06)
4 = Frequently Plains
5 = Very frequently
Unable to get emotionally close (SRK29f) 1 = Never 618 Total Sample 3.18 (.04)
2 = Infrequently 314 Southwest 3.21 (.05)
3 = Sometimes 304 Northern 3.15 (.06)
4 = Frequently Plains
5 = Very frequently
Afraid to go to sleep (SRK29ff) 1 = Never 619 Total Sample 2.13 (.04)
2 = Rarely 314 Southwest 2.14 (.06)
3 = Sometimes 305 Northern 2.11 (.06)
4 = Frequently Plains
5 = Almost always
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE = standard error
*p <.05, ** p<,01, *** p<.001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics (Cont.)
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Have nightmares of service events (SRK29g) 1 = Never 620 Total Sample 2.75 (.05)
2 = Infrequently 315 Southwest 2.84 (.06)
3 = Sometimes 4 = Frequently 305 Northern Plains 2.66 (.06)
5 = Very frequently
Stay away from military reminders (SRK29gg) 1 = Never 619 Total Sample 2.60 (.04)
2 = Rarely 314 Southwest 2.64 (.07)
3 = Sometimes 305 Northern Plains 2.57 (.07)
4 = Frequently 5 = Almost always
Think of service events/wish dead (SRK29h) 1 = Never true 620 Total Sample 1.90 (.04)
2 = Rarely true 315 Southwest 1.94 (.06)
3 = Sometimes true 305 Northern Plains 1.86 (.06)
4 = Frequently true
5 = Very frequently true
Memory is as good as ever (SRK29hh) 1 Not at all 1 = Almost 619 Total Sample 2.94 (.05)
true always true 314 Southwest 2.79 (.06)
2 = Rarely true 2 = Usually true 305 Northern Plains 3.32 (.06) **
3 = Somewhat 3 = Somewhat
true true
4 = Usually true 4 = Rarely true
5 = Almost always true 5 = Not at all true
Seem to lack feelings (SRK29i) 1 = Never true 620 Total Sample 2.50 (.05)
2 = Rarely true 315 Southwest 2.49 (.07)
3 = Sometimes true 305 Northern Plains 2.51 (.07)
4 = Frequently true
5 = Very frequently true
Hard time expressing feelings (SRK29ii) 1 = Not at all true 619 Total Sample 3.05 (.05)
2 = Rarely true 314 Southwest 3.03 (.06)
3 = Sometimes true 305 Northern Plains 3.08 (.07)
4 = Frequently true
5 = Almost always true
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE = standard error
*p <.05, ** p<-01, *** p<.001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics (Cont.)
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Feels like killing self (SRK29j) 1 = Not at all true 620 Total Sample 1.56 (.04)
2 = Slightly true 315 Southwest 1.63 (.06)
3 = Somewhat true 305 Northern Plains 1.48 (.05)
4 = Very true 5 = Extremely true
Stay asleep until alarm goes off (SRK29k) 1 = Never 1 = Very 618 Total Sample 2.18 (.04)
2 = Infrequently frequently 314 Southwest 2.18 (.06)
3 = Sometimes 2 = Frequently 304 Northern Plains 2.18 (.06)
4 = Frequently 3 = Sometimes
5 = Very 4 = Infrequently
frequently 5 = Never
Not able to remember military experience 1 = Not at all true 618 Total Sample 2.79 (.04)
(SRK29kk) 2 = Slightly true 314 Southwest 2.96 (.06)
3 = Somewhat true 304 Northern Plains 2.62 (.06)**
4 = Very true 5 = Extremely true
Wonder why still alive (SRK291) 1 = Never 620 Total Sample 2.56 (.05)
2 = Infrequently 315 Southwest 2.43 (.07)
3 = Sometimes 4 = Frequently 305 Northern Plains 2.70 (.07)**
5 = Very frequently
Some situations feel like in service (SRK29m) 1 = Never 620 Total Sample 2.54 (.05)
2 = Infrequently 315 Southwest 2.54 (.05)
3 = Sometimes 4 = Frequently 305 Northern Plains 2.54 (.05)
5 = Very frequently
Dreams cause to wake in cold sweat (SRK29n) 1 = Never 620 Total Sample 2.39 (.05)
2 = Infrequently 315 Southwest 2.40 (.07)
3 = Sometimes 4 = Frequently 305 Northern Plains 2.41 (.06)
5 = Very frequently
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE = standard error
*p <.05, ** p<-01, *** p<.001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics (ConO
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Feel like cannot go on (SRK29o) 1 = Not at all true 620 Total Sample 1.86 (.05
2 = Rarely true 315 Southwest 1.87 (.06)
3 = Sometimes true 305 Northern Plains 1.86 (.06)
4 = Very true
5 = Almost always true
Do not laugh or cry at same things (SRK 29p) 1 = Not at all true 620 Total Sample 2.51 (.05)
2 = Slightly true 315 Southwest 2.53 (.07)
3 *> Somewhat true 4 = Very true 5 = Extremely true 305 Northern Plains 2.50 (.07)
Enjoy many things used to enjoy (SRK 29q) 1 Never true 1 = Always true 620 Total Sample 2.58 (.05)
2 = Rarely true 2 = Usually true 315 Southwest 2.53 (.05)
3 = Sometimes 3 = Sometimes 305 Northern Plains 2.52 (.06)
true true
4 = Usually true 4 = Rarely true
5 = Always true 5 = Never true
Daydreams are real/frightening (SRK29r) 1 = Never true 620 Total Sample 2.20 (.05)
2 = Rarely true 315 Southwest 2.26 (.06)
3 = Sometimes true 305 Northern Plains 2.15 (.06)
4 = Frequently true 5 = Very frequently true
Easy to keep a job since military (SRK29s) 1 = Not at all 1 = Extremely 618 Total Sample 3.17 (.05)
true true 314 Southwest 3.16 (.07)
2 = Slightly true 2 = Very rare 304 Northern Plains 3.21 (.07)
3 = Somewhat 3 = Somewhat
true true
4 = Very rare 4 = Slightly true
5 = Extremely 5 = Not at all
true true
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE standard error
*p <.05, ** p<.01, *** p<.001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics (Cont.)
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Trouble concentrating on tasks (SRK 29t) 1 = Never true 619 Total Sample 2.57 (.04) Trouble
2 = Rarely true 314 Southwest 2.63 (.06) concentrating
3 Sometimes 305 Northern 2.51 (.06) on tasks (SRK
true Plains 29t)
4 = Frequently true 5 = Very frequently true
Enjoy company of others (SRK29v) 1 = Never 1 Never 619 Total Sample 2.59 (.04)
2 *= Rarely 2 = Rarely 314 Southwest 2.64 (.05)
3 = Sometimes 3 = Sometimes 305 Northern Plains 2.53 (.06)
4 = Frequently 4 = Frequently
5 = Very 5 = Very
frequently frequently
Frightened by urges (SRK29w) 1 Never 619 Total Sample 2.38 (.04)
2 Rarely 314 Southwest 2.43 (.06)
3 = Sometimes 4 = Frequently 305 Northern Plains 2.34 (.06)
5 = Very frequently
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE standard error
*p <.05, ** p<.01, *** p<.001 for differences between tribes

Table 3.4: MPTSD Scale Item Characteristics
MPTSD Scale Response Set/Recoded Set N Tribe Mean (SE)
Fall asleep easily (SRK29x) 1 = Never 1 Very 619 Total Sample 3.11 (.04)
2 = Rarely frequently 314 Southwest 3.14 (.06)
3 = Sometimes 2 = Frequently 305 Northern Plains 3.08 (.06)
4 = Frequently 3 = Sometimes
5 = Very 4 Rarely
frequently 5 Never
Jump at unexpected noises (SRK29y) 1 =Never 619 Total Sample 3.25 (.05)
2 = Rarely 314 Southwest 3.11 (.06)
3 = Sometimes 4 = Frequently 5 = Very frequently 305 Northern Plains 3.40 (.06)**
No one understands feelings (SRK29z) 1 = Not at all true 619 Total Sample 2.80 (.05)
2 = Slightly true 314 Southwest 2.91 (.07)
3 = Somewhat true 305 Northern Plains 2.70 (.06)*
4 = Very true 5 = Extremely true
Scale items in parentheses correspond with actual items from the AIWP interview
MPTSD = Mississippi Combat-related Post-traumatic Stress Disorder Scale
SE = standard error
*p<-05, ** p<.01, *** p<.001 for differences between tribes

Item Correlations. Table 3.5 provides the results from the correlations with
the Northern Plains above the diagonal and the Southwest below the diagonal. All
values were significant at p<.001 and demonstrated moderate inter-item
correlations and consistent patterns across tribes.
All items were positively correlated and consistent between tribes, although
some of the items had higher correlations than others. The results indicated low to
high correlations among the re-experiencing/avoidance items (SRK29bb, SRK29cc,
SRK29d, SRK29e, SRK29ee, SRK29ff, SRK29g, SRK29gg, SRK29i, SRK29ii,
SRK29kk, SRK291, SRK29m, SRK29n, SRK290, SRK29p, SRK29r, SRK29t,
SRK29w, SRK29y, SRK29z) ranging from .14 to.80 for both tribes. The
correlations among items about lack of control (SRK29aa, SRK29hh, SRK29q,
SRK29s) ranged from .29 to.51 for the Southwest and from .36 to .52 for the
Northern Plains. The arousal (SRK29k, SRK29x) items were also consistent across
tribes with correlations ranging from .26 to .27 for both tribes. The withdrawal and
avoidance (SRK29dd, SRK29f, SRK29v) items were slightly higher for the
Northern Plains, ranging from .30 to .35 compared to .27 to .30 for the Southwest.
The correlations for the (SRK29h, SRR29j) suicidality items were slightly higher,
ranging from .58 to.60 for both tribes. Finally, the correlations for the guilt
(SRK29a, SRK29b) items were lower, ranging from .16 to. 17 for both tribes.

Table 3.5: MPTSD Item Correlations by Tribe
Item SRK29a SRK29aa SRK29b SRK29bb SRK29c SRK29cc SRK29d SRK29dd SRK29e SRK29ee SRK29I SRK29ff
SRK29a 0.18 0.17 0.21 0.18 0.20 0.30 0.07 0.30 0.17 0.07 0.24
SRK29aa 0.19 0.12 0.11 0.24 0.25 0.28 0.23 0.27 0.26 0.18 0.30
SRK29b 0.16 0.10 0.03 0.03 0.03 0.00 0.00 0.07 0.08 0.12 0.04
SRK29bb 0.23 0.12 0.02 0.32 0.44 0.37 0.15 0.35 0.29 0.06 0.44
SRK29c 0.19 0.25 0.04 0.32 0.36 0.47 0.13 0.45 0.39 0.09 0.33
SRK29cc 0.21 0.26 0.03 0.44 0.36 0.42 0.20 0.39 0.50 0.02 0.53
SRK29d 0.30 0.28 0.00 0.37 0.47 0.41 0.18 0.53 0.45 0.03 0.49
SRK29dd 0.08 0.23 0.01 0.15 0.12 0.21 0.18 0.14 0.11 0.22 0.16
SRK29e 0.31 0.29 0.07 0.36 0.46 0.39 0.52 0.14 0.46 0.04 0.44
SRK29ee 0.18 0.28 0.07 0.30 0.40 0.49 0.44 0.11 0.47 0.11 0.57
SRK29f 0.07 0.18 0.12 0.06 0.09 0.02 0.02 0.21 0.04 0.10 0.03
SRK29ff 0.26 0.31 0.05 0.45 0.35 0.52 0.50 0.17 0.45 0.57 0.03 gjpggipi
SRK29g 0.17 0.21 0.02 0.51 0.46 0.53 0.55 0.21 0.43 0.41 0.03 0.62
SRK29gg 0.23 0.16 0.03 0.55 0.43 0.54 0.58 0.22 0.45 0.43 0.00 0.60
SRK29h 0.23 0.32 0.02 0.40 0.37 0.41 0.50 0.11 0.47 0.39 0.09 0.53
SRK29hh 0.16 0.41 0.03 0.18 0.09 0.15 0.28 0.28 0.21 0.35 0.19 0.34
All values significant at p <.001

Table 3.5: MPTSD Item Correlations by Tribe (ContT
Item SRK29a SRK29aa SRK29b SRK29bb SRK29c SRK29cc SRK29d SRK29dd SRK29e SRK29ee SRK29f SRK29ff
SRK29i 0.23 0.23 0.01 0.43 0.43 0.41 0.48 0.18 0.48 0.45 0.04 0.55
SRK29ii 0.31 0.26 0.02 0.40 0.32 0.40 0.39 0.25 0.43 0.48 0.16 0.46
SRK29j 0.24 0.26 0.00 0.25 0.32 0.33 0.39 0.17 0.41 0.30 0.03 0.41
SRK29k 0.03 0.10 0.08 0.04 0.03 0.02 0.03 0.14 0.11 0.05 0.25 0.02
SRK29kk 0.20 0.22 0.02 0.34 0.26 0.43 0.37 0.21 0.36 0.38 0.02 0.41
SRK29I 0.34 0.20 0.04 0.46 0.36 0.44 0.45 0.12 0.44 0.38 0.01 0.44
SRK29m 0.21 0.21 0.02 0.40 0.35 0.42 0.45 0.09 0.46 0.47 0.05 0.50
SRK29n 0.21 0.29 0.03 0.52 0.38 0.56 0.53 0.17 0.45 0.44 0.02 0.65
SRK29o 0.28 0.34 0.04 0.43 0.35 0.50 0.49 0.15 0.48 0.46 0.01 0.63
SRK29p 0.27 0.19 0.11 0.42 0.37 0.42 0.42 0.22 0.47 0.42 0.03 0.51
SRK29q 0.24 0.51 0.05 0.15 0.21 0.25 0.32 0.18 0.27 0.25 0.26 0.33
SRK29r 0.18 0.19 0.02 0.39 0.32 0.42 0.42 0.14 0.34 0.40 0.04 0.54
SRK29s 0.08 0.31 0.05 0.08 0.19 0.24 0.24 0.15 0.21 0.31 0.17 0.25
SRK29t 0.23 0.24 0.05 0.35 0.40 0.42 0.48 0.16 0.42 0.44 0.02 0.51
SRK29v 0.29 0.43 0.05 0.24 0.24 0.25 0.27 0.35 0.25 0.30 0.30 0.36
SRK29w 0.22 0.20 0.08 0.41 0.35 0.44 0.50 0.04 0.43 0.52 0.03 0.57
SRK29x 0.18 0.33 0.01 0.23 0.15 0.31 0.28 0.26 0.19 0.25 0.12 0.47
SRK29y 0.22 0.22 0.06 0.38 0.41 0.43 0.51 0.14 0.40 0.42 0.01 0.52
SRK29z 0.33 0.26 0.09 0.47 0.34 0.52 0.45 0.23 0.49 0.47 0.12 0.52
All values significant at p <.001

Table 3.5: MPTSD Scale Item Correlations by Tribe (ConO
Item SRK29g SRK29gg SRK29H SRK29hh SRK29i SRK29ii SRK29j SRK29k SRK29kk SRK291 SRK29m SRK29n
SRK29a 0.16 0.21 0.21 0.16 0.21 0.31 0.22 0.04 0.20 0.33 0.20 0.20
SRK29aa 0.21 0.15 0.30 0.42 0.21 0.24 0.24 0.11 0.22 0.18 0.20 0.28
SRK29b 0.03 0.02 0.02 0.02 0.01 0.04 0.00 0.07 0.03 0.04 0.04 0.01
SRK29bb 0.52 0.56 0.39 0.17 0.42 0.40 0.24 0.04 0.34 0.46 0.40 0.53
SRK29e 0.46 0.43 0.36 0.09 0.42 0.32 0.32 0.02 0.27 0.35 0.34 0.38
SRK29cc 0.54 0.54 0.41 0.16 0.40 0.39 0.32 0.00 0.44 0.43 0.41 0.56
SRK29d 0.55 0.59 0.51 0.29 0.48 0.39 0.40 0.03 0.37 0.46 0.46 0.54
SRK29dd 0.21 0.22 0.09 0.29 0.17 0.25 0.15 0.13 0.20 0.12 0.08 0.16
SRK29e 0.42 0.45 0.46 0.21 0.47 0.42 0.39 0.11 0.36 0.42 0.45 0.44
SRK29ee 0.40 0.44 0.37 0.34 0.44 0.48 0.28 0.05 0.37 0.36 0.45 0.43
SRK29f 0.04 0.00 0.09 0.20 0.03 0.17 0.02 0.24 0.03 0.01 0.05 0.02
SRK29ff 0.61 0.60 0.52 0.34 0.54 0.47 0.39 0.02 0.42 0.44 0.51 0.66
SRK29g 0.64 0.59 0.19 0.51 0.35 0.40 0.01 0.49 0.52 0.55 0.80
SRK29gg 0.63 0.51 0.22 0.52 0.47 0.36 0.03 0.46 0.53 0.50 0.62
SRK29h 0.58 0.50 0.22 0.56 0.30 0.58 0.14 0.39 0.52 0.47 0.60
SRK29hh 0.19 0.20 0.23 0.22 0.33 0.18 0.05 0.36 0.21 0.13 0.22
All values significant at p <.001

Table 3.5: MPTSD Scale Item Correlations by Tribe (Cont.)
Item SRK29g SRK29gg SRK29h SRK29hh SRK291 SRK29ii SRK29J SRK29k SRK29kk SRK291 SRK29m SRK29n
SRK29i 0.51 0.52 0.57 0.21 0.45 0.40 0.13 0.37 0.50 0.49 0.53
SRK29ii 0.34 0.47 0.31 0.32 0.45 0.22 0.04 0.44 0.39 0.44 0.39
SRK29j 0.41 0.36 0.60 0.18 0.42 0.23 0.10 0.28 0.40 0.29 0.46
SRK29k 0.00 0.03 0.14 0.07 0.12 0.05 0.10 0.16 0.09 0.12 0.01
SRK29kk 0.47 0.43 0.40 0.37 0.36 0.43 0.29 0.17 0.36 0.39 0.47
SRK29I 0.52 0.52 0.52 0.22 0.51 0.39 0.42 0.10 0.36 hISHP 0.53 0.58
SRK29m 0.55 0.48 0.48 0.14 0.50 0.43 0.31 0.13 0.40 0.54 0.58
SRK29n 0.80 0.61 0.60 0.22 0.54 0.38 0.47 0.02 0.46 0.58 0.59
SRK29o 0.60 0.51 0.68 0.28 0.60 0.42 0.60 0.13 0.45 0.56 0.53 0.71
SRK29p 0.50 0.50 0.45 0.26 0.55 0.45 0.36 0.09 0.44 0.49 0.46 0.53
SRK29q 0.23 0.22 0.29 0.34 0.25 0.26 0.24 0.09 0.25 0.24 0.21 0.28
SRK29r 0.51 0.49 0.47 0.21 0.51 0.34 0.37 0.02 0.42 0.43 0.47 0.53
SRK29s 0.16 0.18 0.23 0.28 0.16 0.26 0.20 0.16 0.11 0.15 0.19 0.23
SRK29t 0.46 0.50 0.46 0.33 0.52 0.51 0.35 0.15 0.51 0.49 0.55 0.48
SRK29v 0.24 0.27 0.28 0.33 0.30 0.28 0.24 0.12 0.26 0.25 0.17 0.29
SRK29w 0.46 0.49 0.48 0.34 0.49 0.40 0.37 0.16 0.42 0.45 0.44 0.51
SRK29x 0.34 0.29 0.30 0.26 0.27 0.21 0.30 0.26 0.19 0.23 0.19 0.41
SRK29y 0.54 0.50 0.40 0.24 0.38 0.43 0.24 0.00 0.37 0.43 0.46 0.57
SRK29z 0.46 0.56 0.46 0.32 0.51 0.56 0.40 0.10 0.48 0.44 0.43 0.47
AH values significant at p <.001